Visual Information Design

From CS160 Spring 2013
Jump to: navigation, search



Resources for Design Patterns

Reading Responses

Soo Hyoung Cheong - 4/1/2013 19:37:30


1) This visualization is based upon the data of the Earthquakes that occurred in the world in the past 7 days. The "data points" are based on the number of days ago it occurred (categorized by the color of points), the location of the epicenter, and magnitude of the earthquake (indicated by the radius of point centered on the epicenter). The number of days ago the earthquake occurred (the color coding) is Quantitative TIme variable, the location of the epicenter is the quantitative geographical variable (which is determined by coordinates), and the magnitude of the earthquake is quantitative spatial variable (which by the size of the point show the magnitude of earthquake).

2) It used a world map with 2 dimensional representation of points/coordinates (in another words, circles). It also shows lines that indicate plate boundaries and faults.

3) This visual structure is very effective because it clearly shows the location of the earthquakes and how strong the earthquake was in just a glimpse. If these visualized data were just given in numbers and coordinates, it would not mean too much for the viewer, since it is hard to find patterns through comparison analysis between every pair of points. Instead, this visual structure shows all those information and patterns that are easily recognizable in one bird-eye view.

Brian L. Chang - 4/1/2013 23:18:35

The visualization is based on the internet traffic around the world. The data is taken and then the average is computed. From there the map shows hotspots and areas that have above average network traffic. The data is based on bytes (although it doesn't matter what unit is used since it is converted to a percentage). The monitor uses a map of the world that is colored according to the network traffic. The colors range from green to red with green being normal or below usage and red being the highest usage. There are intermediate colors to show values in between. The chosen visualization is effective because it takes in a lot of data and presents it in a way that we can understand quickly. The colors range like many heat maps with green being cool and red being hot. This allows users to instantly recognize which areas have heavy (or hot) internet traffic. The map allows users to see where the traffic is and to see how it relates to other areas rather than there just being a country/state name attached to a number.

Tiffany Jianto - 4/2/2013 10:48:20

1) what kind of data tables the visualization is based on, including variable types The data tables that the visualization is based on are information from Facebook and Facebook graph to find connections with and among your friends based on their geographic location which is mapped to over a globe. The different variables that are used to categorize your friends and their locations are “Hometown,” “Current,” “Education,” and “Workplace.” Based on these variables, links across a 3D globe are created to map out all your friends’ locations, as well as connect people who share the same location from the variable selected. For example, if "Current" is selected, lines reach out to connect the different geographic locations across the globe of places your friends are currently in to display the location and to show all you friends who currently live in this location.

2) what kind of visual structures are used The visual structure used in an interactive 3D globe. When a variable to categorize by is selected, lines that connect your current location to the geographic location of your friends are drawn out over the globe. From here, you can click on any of those lines which will take you to the location selected and display which friends share this location based on the category selected.

3) why the chosen visualization is effective. This chosen visualization is effective because it is over a 3D globe, which is something that everyone is familiar with. Furthermore, the globe is able to turn and move with the user’s mouse click and drags. The visualization is familiar to the user, straightforward, and easy to use. It makes a clear mapping of geographic location to friends.

Include a URL to the visualization. The link is to the home screen / beginning of the application. To look at more information, graphs, and details, use the arrow keys (hitting right will take you towards the visualization itself)

Ryan Rho - 4/2/2013 20:35:20

PadMapper PadMapper is a online service that finds you apartments in a certain location that meets your criteria. You could search for apartments that are within 15 miles from your office and cost less than $1500 per month per person.

1) what kind of data tables the visualization is based on, including variable types

Variable types are the factors of apartments. The variables can be change in the options of the website. Common variables are the geographical location of an apartment, rental cost, distance from a specific location (i.e. your office), and number of bedrooms. Most of the variables are comparable, so they are ordinal. Even the geographical location is ordinal because some would prefer a certain location. Apartment names and availability of certain factors (usually binary information such as 'cats allowed') are nominal.

2) what kind of visual structures are used

The data table is mapped into a geographical map where the location of each apartment is the coordinate. Technically, it is a scattered plot where the x-axis is latitude and y-axis is longitude.

3) why the chosen visualization is effective.

This visualization is effective in that regardless of the number of factors of choosing an apartment, the visualization keeps it simple. Although other factors may be important and some of them are more important to others for some users, this visualization tries to prioritise the factors. Since most users think the location is important, the visualization is a map rather than a list. After that the user can click each apartment and check other factors such as number of bedrooms.

Annie (Eun Sun) Shin - 4/2/2013 21:50:45 The above link contains an interesting information visualization I saw during the most recent presidential campaign. Titled "512 Paths to the White House," the visualization allows users to select a winner in the most competitive states below to see all the paths to victory available for Obama or Romney. This information visualization made by the New York Times is an online, interactive, visual representation of abstract data that amplifies cognition. It helps people understand all the possible scenarios and resulting outcomes. The kind of data table the visualization is based on is similar to table 1.10 on page 19 of the reading. The visual structure used is a binary tree which can be represented by a data table with nodes as columns and edges/links and winner as rows. In a data table, the rows are considered variables. The values in each element of the edge/link row would be the children of the node (can be 0, one other node, or two other nodes, while the values in each element of the winner row will be set to Obama or Romney. The chosen visualization is effective because it takes data and maps it into a tree that is easy to comprehend and interact with. From looking at the visualization, users can quickly find out various information that would have otherwise taken much longer to figure out with just data alone. The visualization provides a different perspective on raw data and organizes it efficiently and aesthetically. The visualization's interactive functionality also helps users filter out data presented by the visualization, making it even more efficient.

Brent Batas - 4/2/2013 22:18:05

I’m going to discuss Google Finance as an example of compelling information visualization. In particular, I’m going to analyze the visualizations for a company’s page (for the purposes of picking something specific, I’ll pick Microsoft)


1) The data tables that the visualizations are based on are stock price (as a decimal number of some currency, like USD) measured every minute of the day, as well as volume (in mil / 1wk).

2) There are two visual structures on the page of a given company.

First is a line graph representing the company’s stock price as it varies over some time interval. The time interval can be switched from as limited as one day to as broad as ten years, or further, the company’s entire existence. You can also directly compare a company’s stock graph with that of another company’s by entering the other company in the “Compare” field. This overlays the two graphs on the same plane.

Second is a histogram representing the company’s volume (in mil / 1wk) over time.


The line graph of the stock price is effective in several ways. The continuous line illustrates the continuous changes of stock prices, while also making peaks and valleys very clear. It’s easy to see how stock prices fluctuate wildly over day or week intervals, while over a longer interval, general upward, downward, or level trends can be observed. These trends could be useful for figuring out things like how successful a product is, how much people trust the company, or how the economy as a whole is doing.

The line graph makes it easy to notice trends that would be really difficult to notice just by looking at data tables. Looking at the data tables alone, it’s easy to be misled in noticing “false” trends because you’re only looking at a small time interval. And if you do go through the trouble of looking through data tables over a large time interval, it will probably take you a very long time. In addition, the line graph is useful because a line takes up very little space. This allows you to overlay multiple line graphs (different companies’ stock prices) and still have everything be readable.

The histogram of the company’s volume is effective, too, because unlike stock prices, volume is typically meaningful when measured over a week span. This makes the histogram more appropriate since it clearly delimits week intervals at which volume is measured. Restricting the visualization to week-long intervals helps reveal where real “peaks” are in terms of volume, rather than being misled by the wild changes that happen over smaller intervals.

Cory Chen - 4/2/2013 23:59:03

1) In data table form, you would basically have the cut of the meat (rib, sirloin, etc) as one variable and the location on the cow (neck, butt, to the right of the neck, etc) as the other variable. You could also have a more detailed data table that also includes the detailed version of the cuts (rib roast, rib steak, back ribs, etc). 2) The image uses a physical mapping in order to help the user associate the cuts of meat to where they are actually located on the cow. The information could have very easily been presented as just text, which would be much harder to process and remember. The more detailed cuts of meat are also drawn to help the user remember what the names of them are and what they look like. By having all the information be physical and map to a specific location on the cow, the user can just remember the general sections of the cow and know what meat they are getting. 3) The viewer wants to know a mapping of different types of meat to its place on the cow, and the visualization creates a direct mapping of the meat types. The data is obvious and intuitive, and the physical mapping helps the viewer retain the information better because it is in word and picture form. The data doesn't deal with numbers, so representations like a bar graph or a pie chart would not be appropriate.


Elise McCallum - 4/3/2013 1:08:33

I chose to use this visualization of Part One of Jack Kerouac's On the Road, located here: Literary-Organism-Poster.jpg

1) This visualization is based on less of a strict data table and is instead based on word counts in sentences, chapters, and parts of a given novel, On the Road. The other variable type taken into consideration (besides the numerical count of words) is a nominal variable, namely, which character or theme is associated with that word/phrase/sentence. 2) Lines and fan-like structures are the main visual structures used in this information visualization. As a whole, the visual structure of a diagram is also used as a way to display information ( similar to sentence diagramming, but on a more abstracted level). Color is another visual structure implemented to separate and identify information (namely, which character/theme is being addressed) 3) This visualization is effective largely because of its structure. As detailed in the key, the lines represent the divisions of the text into parts, chapters, paragraphs, sentences, and words. Beyond those divisions, pieces of the visualization are colored based on who is speaking/from whose view the dialogue is generated. Examining this visualization, one can see that this structure is effective because it not only allows viewers to see how a book is split apart, but it also allows viewers to see in a different light how a story progresses. One can see which characters play a main role in the beginning, middle, and end, and get a general feel for which parts of the book are lengthy and which are short.

Sihyun Park - 4/3/2013 2:57:45

Politify: Politify is a web application that allows users to see the impacts of each 2012 presidential candidate’s policies on their households.

1. The visualization is based on a data table that maps input variables: Filing status, income, age, zip code, and most importantly, the candidate, to the outputs: Net changes to income, tax benefits, and federal services. Of input variables, income, age, number of dependents, and number of students in college are nominal, zip code is ordinal, and filing status and the candidate is nominal. All of the output variables are quantitative. What's interesting is that the candidate is in essence a package of many different variables (the candidate's policy) that are largely nominal. However, the output is all quantitative variable, thereby converting a qualitative policy into a quantitative data that one can assess the monetary benefits.

2. Politify's visualization uses perceptual and spatial visual structures. It highlights a positive elements as green and negative elements as red, and breaks each sector of tax benefits in different sizes according to its actual amount of monetary benefit. (in number of dollars) In the local section, users can notice that almost everywhere on the map is marked as blue, showing that Obama's policies has a positive effect on more parts of the US in terms of pure "area covered." This is somewhat eluding, as people live in concentrated areas like cities, that might have a strong support for Romney.

3. The visualization is effective, as it allows the user to see how a government policy can affect their lives easily by looking at graphs indicating an immediate monetary benefit to the household. By mapping the input variables along with the candidates' policies to a quantitative output, the user can assess the effectiveness of each candidate's policy without reading through a bunch of text.

Cong Chen - 4/3/2013 3:05:56

I examined the data visualization diagram of "Gun deaths since Sandy Hook" at the URL: (You must scroll down to see the graph)

1) The data from the data visualization graph is based off a tuple data table, where the coordinates of a location in America maps to the number of gun deaths. Essentially, the data table is a 3-D table with 3-axises, longitude, latitude, and gun deaths. The variable type is ordinal because they are specific numbers that map to certain dependent information like gun deaths.

2) This graph uses space and marks. It effectively uses space by having a diagram of the United Stations and effectively encodes the information of the independent variables, the longitude and latitude information (essentially the location in America) with the dependent variable, the number of gun deaths by color. The darker the color, the more number of deaths the visual display represents. The Marks used to show these "visible" things in space are points which essentially form together into 2D circular areas. The information of gun deaths by location is effectively portrayed by this visual marking. Also, the darker the color, the more gun deaths represented.

3) The chosen visualization is effective because it captures the most important information and mapping that users care about when interested in such information. Users are probably most interested in which areas of America have the most gun deaths and which areas do not. By visually marking the information on a map of America, it is logical and easy for a user to tell which areas have more gun deaths because of the markings on the map. The markings take it to the next step by having a gradient, giving user's more information about relatively how many more or less gun deaths are in certain areas. The graph fails to account for population density and show this information to the user so that users can account for this when considering the gun death information. However, ultimately, the graph visualization is quite effective because it really captures the most important part: showing the relationship between location and number of gun deaths.

Matthew - 4/3/2013 11:23:51

Visualization: The-World-Votes-Infographic.png

This visualization is based on survey results conducted in Brazil, Colombia, India, Indonesia, Kenya, Nigeria, Pakistan, Philippines, and Vietnam. There were 10 different questions asked each with multiple choice answers. The data is from 1900 respondents in those countries.

There are multiple visual structures in use here. The first are the use of bar graphs. They provide easy comparison and the clear demarcation of the value of each bar makes it easy to compare between the various candidates. Also, in the breakdown, there is usage of grouped histograms, which allow for easy comparison of political opinion between the countries for various countries. Here, the use of a restricted color palette emphasizes some of the columns, helping to draw the eye toward it. Pie graphs are also crucial to visualization. They aide in making relative comparison between the candidates and once again draw the eye to particular features. An interesting visualization is used in 10, where the picture of each of the US leaders' pictures are scaled based on their percentage, which reinforces the concept of bigger means more important.

Sumer Joshi - 4/3/2013 12:04:22


I chose to analyze/look at a Facebook Social Graph Visualization. I thought that this was pretty cool/effective because you can see how people's friendships are interconnected with each other, which mutual friends they might know, and finally, how other parts of their life (School, Work) might be linked together.

The visualization is based off of real world actions and objects from the Facebook Graph API, and the variable types in this case are users (users in the network), actions (what actions a person I might take), and an actor (someone who takes that action)

The visual structure that is used is graph that is connected with circles that display users, and it is the simplest way to do so.

Jin Ryu - 4/3/2013 12:11:38

One compelling information visualization online is "Google Maps". It displays geographical data about the Earth and is interactive. URL is:

1. The data tables the visualization is based on are:

  • locations [variables and types: names (of cities/bodies of water/places of interest) - nominal; type of place (shop, business, museum, park, garden, etc) - ordinal; addresses - nominal; city/country - ordinal; zipcode - quantitative; latitude/longitude - quantitative; expanse of area/coordinate points - quantitative; color on map - ordinal; diagram/picture - nominal]
  • photos [variables and types: location - nominal; latitude/longitude - quantitative; actual image - nominal]
  • paths/streets [variables and types: names (of streets, highways) - nominal; intersections of paths/streets - nominal; type of pathway (local, freeway, highway, trail, etc.) - ordinal; addresses - nominal; city/country - ordinal; zipcode - quantitative; latitude/longitude - quantitative; length of path/coordinate points/lines - quantitative; color on map - ordinal; diagram/picture - nominal]
  • directions [connecting paths/streets - nominal; location place A and location place B - nominal; distance - quantitative; time (how long it takes to get there using estimated velocity and distance) - quantitative]
  • distance [start city - nominal, end city - nominal, distance - quantitative]
  • scale [kilometers or miles per pixel/inch in map view - quantitative]
  • traffic [traffic report - nominal; congestion intensity - possibly quantitative by estimated time delay]

2. Visual structures used are:

  • colored maps, or visual representation of area using colors, shapes, and lines; also shows physical geographical features such as mountains using different texture/patterns/shading
  • diagrams for buildings and streets in map view that are colored: paths or small streets (white), large streets (yellow), highways or freeways (orange), gardens, forests, and parks (green), water (blue), buildings or land (tan), campus (brown), etc.
  • actual photographs put together as a 3-D panorama at ground level for street view
  • satellite images taken from above [aerial view]
  • icons (green trees for parks and gardens, blue/white shields for large freeways, white ovals for smaller highways, spoon and fork for restaurants, bed for bed and breakfasts, hats for academia, transit icons, etc.)

3. The visualization is very effective because:

  • it is interactive and can be directly manipulated by these: modifiable view since it can zoom in and out; map image is panning; there are clickable places (more detail/information, relevant photos, review, etc.)
  • switch from map view to street view and back
  • there is a search query which makes finding locations or places of interest easier
  • getting directions is immediately visual on map and easy to understand [only need to designated point A and point B which retrieves a possible route and its alternatives; marks points A and B with green bubbles, then colors this route blue on map as you select a version of suggested route]
  • finding user's current location (orients the user visually onto the map)
  • all places and streets/paths are named or have some kind of label to identify what they are; sometimes further details are included in a bubble-popup when clicked
  • shapes are diagrammed relatively accurate on map (ex: building images outlined to match area and shape)
  • intuitive spatial/geographic indicators such as ridge/groove marks to indicate changing elevation (like mountains) or blue for water and green for foresty areas

Haotian Wang - 4/3/2013 12:35:52

I chose the standard example of a stock-market chart, which i feel is great information visualization even though it's very old.

1) The data tables is ultimately based on stock prices, so the variable type is monetary value. Different data tables are created with different mathematical functions based on the prices, so that the actual price-chart is differently visualized than each of the technical indicators. 2) Visual structure that is used is a candlestick-graph for the main chart, which visualizes the price changes within each time period by each candle's relative heights. This is effectively a line-graph with candles instead of line-points, since candles are able to visualize more features such as day-start price and day-end price. The technical indicators (at least for this example) are also effectively line graphs, which the line height indicating the value of some math function of the last X days of prices. 3) The main candle-stick price chart is very effective because it's both simple and conveys price patterns very well. Many books have been written of chart analysis because the visual patterns in the chart convey inter-day trends in pricing. Trends including an upward-trend, which is just the candlesticks going up day by day, a trading range, which is the candlesticks fluctuating within a certain height range, and even a head-and-shoulders, with the chart having a small mountain-shape, then a larger mountain-shape right away, then another smaller mountain-shape, making the appearance of "head and shoulders". This kind of pattern recognition (especially head and shoulders) would be very difficult to deduce from simple raw price data, but takes no more than a second to recognize when put into chart form. Thus it really speeds up problem solving for the task of recognizing patterns and thus making a trade based on those patterns.

kayvan najafzadeh - 4/3/2013 12:35:56

I chose web interface which is a very simple website to test your internet connection speed. The visualization of first map is based on a data table with many servers around the world, their location on the map, information regarding each one, and their IP address. The visual structure uses a map with my location and other dots (which each represent a server) visible on the map. by hovering over the dots we get detailed regarding that specific server. Because a map is the best representation of the world and by plotting servers on the map, the distance between my location and the targeted server become visible as well.

Minhaj Khan - 4/3/2013 13:05:32

url: 2.png

This visualization is based on comparisons between mac and pc users. the data and variables include geographics, demographics, personality types, preferences, etc, and are measured in comparative percentages. each variable is measured for both mac and pc users and compared.

the visual structure is laid out in 2 long columns, one each for pc and mac users, and flows from top to bottom in order of categories of comparisons. there are various icons and graphics representing the topic of each comparison, as well as a horizontal bar comparing percentages at the top. The comparisons are sectioned off by category horizontally. The information is primarily in text and numbers aside from the organization of information.

This visualization is effective because it compares and contrasts the two types of users directly side by side. All information is laid out in categories for the user to be able to organize information mentally. Various colors and icons are used to engage both right and left brain activity, making the perception of information presented more pleasing.

Avneesh Kohli - 4/3/2013 13:21:12


The visualisation of data tables is based on data that reflects the current NBA player statistics for the 2012-2013 season.The data is ordinal, as the numbers are clearly meant to suggest which players are better based on having higher statistics in various categories. This visualization is particularly effective because it allows you to sort it by any column in the table. It also does a great job of reflecting the ordering and suggesting which of the categories are important to look at by their ordering of the columns. There aren't any visual structures visible.

Thomas Yun - 4/3/2013 13:27:19

The type of visualization that I will be talking about is called a Web Trend Map

This Web Trend Map plots big/well known internet websites onto the Tokyo Metro Map. The different variables include the site itself, rankings on stability and success, and the type of trend. Stability is measured by how wide the stacks look whereas success is determined by height. Every site is also connected by a colored line (similar to train maps) in which each colored line represents a different trend. In addition, people who are well known and associated with a certain site are shown on the map as well. Lastly, the position of the site is accurately placed according to real Tokyo locations. For example, google and its network are found around a certain district in Tokyo that's highly popular or busy because Google sites share the same characteristics. The map itself blends different type of graphs that include things like 3D bar graphs and line graphs. The visualization is fairly useful because a person can instantly see what sites are fairly popular according to trend. It is especially useful to those that know different locations in Tokyo and where each location maps to on the map as it allows users to instantly know which "area" of sites may fit their interests.

Timothy Ko - 4/3/2013 13:32:20

I chose a map of all the lights in the world, which was a visualization produced from data gathered from a satellite launched in 2011. More information can be found in the url at the bottom.

The visualization seems to be based on an input/output data table. Where the input is actually made up of two variables, latitude and longitude. The output is one variable: whether there is a light at the point specified by the latitude and longitude. The latitude and longitude variables are quantitative types. The isLit (as we’ll call it for convenience) output variable is a nominal type that takes on one of two possible values, true or false.

2) what kind of visual structures are used

The visual structure of a world map is used to give context to the data. While the representation of countries isn’t mentioned anywhere in the data table, without this the data would be much more difficult to interpret. The visual structure here works like this: for every class, a white dot is painted on the given longitude/latitude coordinate if isLit is true. Otherwise, no white dot is painted. When enough white dots are painted in a concentrated area, this is a good indication that there is a lot of light in the area, probably a city or group of cities. This is explained in the given url as well.

The chosen visualization is effective because not only can you easily see the many thousands of cases from the data table in just one picture, but you can also easily see relations between the points, as well as make interesting observations. For example, you can see that consumption of electricity is much higher in the eastern half of the United States than in the west.

Ben Dong - 4/3/2013 13:50:38

Using the "states sized by number of electoral votes" graphic at:

1) The data table contains the state name (nominal), the number of electoral votes for that state (quantitative), and the party affiliation based on polls (ordinal).

2) The visual structure is a map visualization where states are sized by the number of electoral votes they have. It maps state size to number of electoral votes (and thus significance in the election) and color to party affiliation.

3) The visualization is effective because it provides an easy way for viewers to quickly understand the current state of the election polls. By using state size to represent electoral vote count, the visualization allows viewers to easily see who is in the lead, as well as which states are most important. The party colors make it easy to see which candidate has which states, and all of the states are in their approximate geographic locations, which allows for easier direct mapping in that regard.

As a side note, clicking through the "next" button transforms the visualization into two large circles (one each for Obama and Romney), placing states into each circle based on who is likely to win that state. States retain their size mapping based on electoral vote count. The fluid animations between these transitions allow viewers to see which states go where, and the relative sizes of the states within each camp (as well as the total size of all the states in any one camp) are great for seeing which candidate is most likely to win in any particular scenario.

yunrui zhang - 4/3/2013 13:53:46

The online example of a compelling information visualization is a pie chart for Mammal species. The url is below:

I will describe it as followings: 1)The data table iis based on Distribution of extant and recently extinct mammal species across orders based on Wilson and Reeder,2005.The rows are mammal spices a nominal variable, and it contains a single column, distributions, a quantitative variable. 2) The visual structure is a pie chart with sections of different sizes and colors. 3) The chosen visualization is effective because a pie chart a in the shape of a circle, and it represents the notion of "100%". Since what we want is to represent distributions that sum up to 100%, a pie chart is ideal because all the data in the data table also sums up to 100%. A human can see a circle within one sight, and it is very fast for a human to interpret if one section of the pie is larger than the other, if they are colored differently. There also many colors available to use, so color the subsections with different colors is also not a problem, and different colors convey more distinctions.

Raymond Lin - 4/3/2013 13:57:00

The Information Visualization I chose was a pretty simple heat map that you might see if you watch the news on tv.

1) In terms of data tables, the visualization is a a forecast on the high temperatures in regions around the United States. These kinds of variables are ordinal variables are they are estimates of the highest temperatures (a < relation).

2) The visual structure used is a spectrum of colors that don't necessarily follow the normal intensity of the light spectrum, but red being higher temps, and blue being lower temps.

3) I think they're effective because once the user understands the layout of the colors, it's simple and very effective way of quickly determining whether it's going to be a hot or cold day. Numbers are too technical and don't necessarily translate as well as a direct comparison with something simple like color.

Oulun Zhao - 4/3/2013 14:02:53

1) The visualization is based on the people’s votes for presidency sorted and categorized by states. The vertical titles of the table could be Obama and Romney and the horizontal titles of the table could be the names of the states. The variable types are quantitative variables because we can do arithmetic on number of votes.

2) The type of visual structures used is spatial substrate because we can see the number of votes sorted into different states. When we mouse over the state we can see the exact number of votes for the two candidates and the same time we can see the general color of the state showing which candidate does that state supports.

3) This visualization is effective because people are familiar with the American map. In addition, it can show clearly which states are supporting Obama and which states are supporting Romney.


Lauren Fratamico - 4/3/2013 14:03:32

I am analyzing the way usgs chooses to represent the most recent earthquaked that have happened on the globe:

They use a map with points on it to show where the earthquake occurred. The points are all circular, but have different sizes and colors. Larger means that it was an earthquake with larger magnitude, and the colors represent how recently it occurred. You can also click on the dots to get more information about that quake. This visualization shows an abstract data type of quantitative data. Each row in the data table is an earthquake that happened recently which includes information about its location, size, and occurrence time. This visualization is effective because it quickly allows you to asses where the most recent earthquakes are and assess their size. It is also interactive, allowing you to find more information about the data if desired.

Elizabeth Hartoog - 4/3/2013 14:03:58

This is an interesting visualization since it's updated dyanmically but still really useful. It goes above and beyond what google maps provides for traffic monitoring for certain areas.

This map provides information on on accidents and shows them as diamonds in different colors. The related variables to the diamonds are: location (assumed lat/long so quantitative), nearest exit (nominal), current speed on highway (quantitative), actual accident (nominal), degree of accident (green/yellow/red...ordinal). This data is most likely pulled from some kind of traffic feed and processed from these variables to be put on the map. The map makes liberal use of points and lines seeing as it's a map. More importantly the accident mapping makes use of points to easily identify the accidents, while also using color to make it easy to spot the more important or traffic heavy accidents through the green <-> red spectrum. While the system is simplistic, putting traffic information onto a traffic map, it is much simpler than a driver compiling the information themselves from online sources/media (quite frankly I wished it worked in the SF bay). It doesn't do anything mind blowing, but it is effective because it is very easy to detect patterns through a colored traffic map and very clearly shows variable of interests (traffic flow/traffic accidents).

Weishu Xu - 4/3/2013 14:07:42

Web Trend Map 4: wtm4-final.png

1) This visualization tracks data on the "Internet's [333] leading names and domains" in order to portray consumer behavior on the Web. The variables that are tracked and processed to be displayed include traffic, revenue, and character. Closely related sites are placed closer to each other in area and metro lines. It also keeps track of the Internet's 111 most influential people.

2) The visual structures used include the Tokyo Metro Line Map as well as "buildings," whose heights would correspond with the site's web presence. The influential individuals are also placed on the map corresponding to where they are likely to be found on the Internet.

3) The chosen visualization is effective because it is very visually appealing and interesting to look at. It allows the viewer to easily compare "presence" between different sites without having to look at extensive tables of tracking several figures. It also helps the user easily identify how websites may be related and how websites may interact based on usage.

Alice Huynh - 4/3/2013 14:25:22

I really like the idea that crazyegg has. What they do is give page analytics in a visual manner by using another type of data visualization "heat graphs" to help web-page owners where the users focus the most on any given web-page.

Crazy Egg relies on the correlation between eye movement and mouse clicking/movement.

The heat map is a type of "diagram" described in the reading as non-interactive. The reading specifies that visualization allows people to use "perception to amplify cognition". The heat map from Crazy Egg uses the heat map technology to symbolize to the designer where the page is be utilized the most so that if there are any problems in that area of the page it would hurt business greatly.

Another really cool visualization that CrazyEgg uses is the "Confetti Tool" that places where people are clicking to a designer's current site. By placing text next to where users are clicking it gives the designer a visualization of where users are coming for and why they are coming their site for.

Zeeshan Javed - 4/3/2013 14:27:47

Find an online example of a compelling information visualization. Describe:

1) what kind of data tables the visualization is based on, including variable types 2) what kind of visual structures are used 3) why the chosen visualization is effective.

The data table the visualization is based on is that of deaths that occurred in the 20th century that were recorded by government infographic data. Information visualization and chart data is used as the visual structures for this particular example. The authors use the size of the bubbles to help the user put in perspective how devastating a particular cause of death is. This way the users can help better compare the cause of deaths in a much more easy to read way. This is much more effective than using raw data and numbers. The visualization also helps categorize major causes of deaths and branches them out into smaller sub categories to point out what is more deadly within a particular category. For example 530 million people die of cancer every year versus the 1.24 million who die of cardiovascular disease. However, the data is made so clear that we can compare items within these large sectors. For example, with effective data visualization we can find that 93 million of the 530 million die of lung cancer as opposed to the 68 million who die of stomach cancer.

Include a URL to the visualization:

André Crabb - 4/3/2013 14:29:22

Example: Github! Every user's Github page shows a graph of their open source contributions.

1. The data is based on commit and push logs over time. Not knowing how Github works, I can imagine that these logs can be made into a spreadsheet that have cells for date, time, code changes, commit message, etc.

2. A 2D matrix graph is used to display the contribution information. The x-axis shows weeks, labeled by months, and the y-axis is days of the week. A full year is displayed in the matrix. Github uses color to show how much was contributed on any specific day of the year. Darker means more contributions, lighter color means less contributions.

3. The chosen visualization is effective because it shows density by colored squares in a matrix, which is fairly easy to understand, especially when the axes are time-related. Also, the use of color is very effective since the darker colors against a light background stand out very well, so its easy to see where the user contributed more work.

URL: (or one with more data:

Alexander Javad - 4/3/2013 14:29:33


"CO2 emissions: This scatter plot was made using Many Eyes, a social data-visualization site. It illustrates the relative amount of carbon dioxide emitted from different countries. This particular visualization has sparked a debate on the politics of the author and the data source itself."

1) Based on CO2 emissions. Variables are CO2 emissions against country.

2) Different sized circles are used to convey the relative size of CO2 emission per country.

3) This is effective because it is very clear which countries are the largest producers of CO2 emissions. You immediately understand the meaning of the data.

Sangyoon Park - 4/3/2013 14:29:54


This website is about earthquake. It shows past earthquakes in the U.S. using several visualization techniques.

1) Navigation charts is used since there is a world map for the view and this can be used to navigate any area that a user wants to view using zoom or drag action. And ordinary grid-style data table is below the map to show lists of details of earthquakes ordered by time it happened. Ordinal variable is used in here since data is organized by time happened in descending order.

2) To help perceptual understanding, there are some different earthquake marks on the map showing how big the earthquake was (bigger earthquake = bigger circle on the map). For the marks, the circle of the point of earthquake shows the exact point where the earthquake happened and how big.

3) For a user who is looking for where/when/how big an earthquake was, this website helps a lot because there is a map that we use almost everyday (considering we use navigation system or google map everyday..), so it is now so much intuitive and easy to understand. Dots(circles) on the map are easily recognizable as earthquakes since they are the only color marks on the map and I assume people who visit this website are mainly looking for earthquakes. Alternatively, grid type data list can help finding specific data if a user is looking for one.

Dennis Li - 4/3/2013 14:30:49

1) The variables are reasons for deaths of humanity, and the number of deaths that those causes incurred. Causes are nominal but the values are quantitative. It is not disclosed what original data tables were used, but the information could have been easily represented with a simple table, where labels would be causes of death and subclasses of those causes, and the entries would be the number of deaths from those causes and sub causes.

2) The author maps the causes of the deaths to circles that represent their significance on humanity. The size of the circle maps to the value of deaths that that cause has incurred. The circles are linked together if they are related and sub causes are always linked to the categorical circle they fall under. What we are left with is a graph of circles.

3) Depending on the size of the circles, the author is able to demonstrate how many people died from a certain cause. This makes it very easy for the audience to see what causes were the greatest factors in deaths of humanity.

Christine Loh - 4/3/2013 15:03:49

1) The visualization is based on a data table that combines relations with metadata that describe it -- this metadata is rows and columns of variables and cases. In this situation, the cases are the Q-values and the variables are the states that result. The relation between them is calculated using the Q-learning equation (taught in CS188!). 2) A 10x10 grid of squares with the Q-value and states visually represented within the grid. 3) The chosen visualization is effective because it successfully takes in the complicated-looking Q-learning equation/formula and turns it into something we can visualize. We can see how the Q-values are affecting each stage, and see how previous Q-values affect later stages. This is particularly effective because we can see cumulative effects rather than just trying to imagine each step in our heads.

Derek Lau - 4/3/2013 15:24:24

Looking at player impact at the bottom of the page:

1. The data tables are based upon selected statistics of each individual player on a given team. The variables are nominal, since the data is a mapping of a number from a player (the nominal variable) to a statistic. 2. Colored bars are used demonstrate the statistical impact of each player to the team. Similarity is the key Gestalt principle used here, as the colors represent one player and grouping the similar colors together gives the viewer an idea of the impact of the player on the team. 3. The chosen visualization is effective because there are two ways to look at the visualization. Per stat (horizontally and by statistic) or per player (vertically and by color). This gives the viewer flexibility in choosing how to assess and analyze the game data and ease in doing so, in whichever way he/she chooses.

Tiffany Lee - 4/3/2013 16:46:31

1) The visualization is based on data tables of crime data which include what kind of crime occurred (nominal), where (nominal) and when (ordinal) it occurred, and the case number of the crime (nominal/ordinal, depending on whether the number assigned randomly or sequentially). 2) The visual structure used is an interactive map with icons placed on the map to show where and what kind of crime occurred; when the user clicks on the icon, more information such as when the crime occurred is shown. The user can also pick the time span of the crimes that show up on the map. 3) This visualization is effective because it allows users to quickly and easily see what kinds of crimes occur where and how often crimes happen in certain areas. It reduces the search and organization time of users to find such information. It also allows for users to quickly see patterns of crime based on location. This visualization groups and frames the information in a way that is useful to people because it focuses on location of crimes which is important information for people that want to be aware of which places are safe or not. Also, the interactive nature of the visualization allows the user to have further control over which information the user wants to see.

Jeffery Butler - 4/3/2013 17:40:33

co-author_space.png 1) The data tables that the visualization is based off of are how many papers the author had published, number of citations they used, and the number of times co-authored. All of these variables are of the type nominal. The data represents how often particular author names occured and in what particular context.

2) The visual structures that are used in this example are Marks and Connection and Enclosure. The number of papers the author published is represented as a circle (larger if there are more papers published). The number of citations is the color of the circle. Lastly, the co-authors are represented as lines draw from node to node creating a connection between authors demonstrating a co-author relationship.

3) The chosen visualization is effective because the larger the node gives the viewer an easy perception of who is the most prominent author. The viewer can also compare the size of a particular node to other authors giving the reader a better understanding of how many more novels the author published than others. The lines running inbetween the nodes tell the user that there is a relationship between these two authors. The darker the lines means that the authors have co-authored together, therefore intutitively, the authors that have a stronger relationship between one another when the line is bolder. The color of the nodes gives the viewer an idea of how much research the author did. The more intense the color the more intense the research (citations). Lastly, the way the nodes are spread across the chart gives the viewer an intuitive sense of who worked with who and who was the most influential in that particular group of authors.

Shujing Zhang - 4/4/2013 1:11:58

1) what kind of data tables the visualization is based on, including variable types:

The visualization is based on data tables for the influence of social sites. The variable are different social websites.

The variable types are nominal because different social websites has no ordinal relationships between each other or quantitative relationships.

2) what kind of visual structures are used

The visual structure that is used is similar to the dot plot histogram. The size of the dot indicates the influence of each of the social website.

3) why the chosen visualization is effective. Include a URL to the visualization.

The URL is at:

The visualization is effective in four ways: First, the websites are located at its corresponding years, so that viewers can find the time they were created in a chronically order. Second, the size of each circle indicates how influential it is. Also, the number in the circle gives concrete measurement of this criterion. From the sizes, it is very easy to compare with other websites. Third, different colors enable viewers to locate circles easily. People can analyze the history of the growth of social network websites as well as forecast the future. Finally, we can also conclude that the general social media marketplace is pretty mature over the past decade.

Mukul Murthy - 4/4/2013 12:49:31

The visualization I chose is

The underlying data tables that this visualization is based on is a table of different things/people. Each item in this table has two attributes: the amount of money that entity spends (or costs, depending on what the entity is), and a category hierarchy that the item falls into. For example, the item Organized Labor's 2011 Campaign Donations falls under the Campaign Donations hierarchy; Steve Ballmer is under the Technology Billionaires hierarchy, which is itself under the Billionaires hierarchy.

Each entity is a nominal data type because there is no natural ordering to them; it doesn't make sense to say a rabbit ownership is any greater or less than dog ownership. Similarly, the hierarchy variable for each entity (its supercategory) is also nominal, because those entries are entities in the data table. However, the amount of money is definitely a quantitative variable; it is continuous and easy to order by amount, and measures such as averages make sense with this type of data.

One of the visual structures is color. Each denomination (one, thousand, million, billion, trillion) is represented by a single square of a different hue (millions is gray, which could be the same hue but very little saturation, but the others are all different hues). This visualization also uses recursion very well, repeatedly partitioning the space into smaller spaces which make sense because they are subcategories of the larger space. It also uses connections and enclosure to separate the different levels, but connect them in the order they are intended to be seen.

The visualization also uses Gestalt principles of organization. The principle of proximity is used to ensure that even though there are 190 unique squares for a bicycle, we group them all together because they are tightly grouped together, and labeled with "Bicycle". Additionally, the Gestalt similarity principle is used in many ways, which include using similar fonts for all levels and similar colors for amounts of the same denomination. We consider one green square in Bicycle the same as a green square in Men's Suit because the green square is the same.

This visualization is overall very effective in its main goal, which is showing the differences in the costs and expenditures of different items. It uses space very well, fitting a ton of information into the graph but still using enclosures and white space to make it look as little cluttered as possible, an impressive feat with so much info. The visualization uses clearly labeled categories so a user always knows what section he is in. Another useful piece is that whenever the denomination changes - for example, thousands to millions - the millions section includes a box with 1000 orange (thousand denomination) squares and shows that it is equal to one gray (million denomination) square, so it is very clear how elements in the millions category compare with those in the thousands category.

Alvin Yuan - 4/4/2013 15:13:07

Visualization URL:, 163413_479288597199_8388607_n.jpg

The data tables behind the visualization have pairs of cities as cases. Variables for each pair of cities included the number of friends between the two cities, the latitudes and longitudes of the cities, and a color value based on the values of the other variables. Each of these variables are quantitative; while color usually isn't used as a quantitative value, in this case color was chosen from a spectrum ranging from black to blue to white, where black indicates few friendships and white indicates lots of friendships. Thus here color has a quantitative aspect tied to it. This visualization uses quantitative axes for the latitude and longitude and composes them orthogonally to create a nice 2D representation of the world. It then makes use of lines with endpoints to represent city pairs. Finally it makes use of a color spectrum applied to the lines to indicate degree of friends between the city pairs. This visualization is effective because it communicates so much with such a small number of variables. There is no notion of continents in the data table (only longitudes and latitudes) and yet the viewer can make them out. It conveys "popular" areas through its use of color, the white areas being the most popular. Finally, it conveys a sense of connectedness by showing just how many connections exist between cities that are continents apart.

Bryan Pine - 4/4/2013 17:51:41

Information Visualization: (the word cloud in the middle of the page).

1. This is a visualization of Sarah Palin's self-introduction as John McCain's running mate. The data table has a nominal variable Word (with a different value for each of many "keywords") and a quantitative variable Count indicating the number of times each word was spoken in the speech. The idea of the word cloud is to visualize this information to allow the user to see at a glance the major topics and direction of the speech.

2. The word cloud is a pretty simple visual structure. The words themselves are displayed (which accounts for the nominal variable) with sizes based on the number of times the word is spoken. According to wikipedia, the size of the words is actually logarithmically based on the number of times that the word is said, which makes it not completely expressive (since the size relationship doesn't directly map the underlying data) but adds to the visual structure's effectiveness because it makes the data easier to interpret by making sure the reader can see everything at once. The structure also leaves out words like "the" that are said all the time because they simply wouldn't be informative to the viewer. Again, this represents a tradeoff of expressiveness for effectiveness.

3. This visualization is effective because it allows someone to get a sense of something dense and nebulous (a speech) at a glance. Normally the reader would have to listen to the whole speech closely and interpret the message of the speaker, but the word cloud works because word count is actually a pretty effective proxy for meaning. Even without listening to the speech, when I look at the map I know that Sarah Palin focused on John McCain, what a great country America is, and the upcoming presidential campaign. Oftentimes, the topic is all I really want to know about a speech, so that might be enough. Of course, this approach does have limits. I can't clearly tell what she said about those topics, and I have to trust that whoever constructed the word cloud chose the words to include and exclude appropriately. Fortunately, those drawbacks don't significantly detract from the purpose of the visualization, which is to show the major topic of the speech at a glance.

Colin Chang - 4/4/2013 19:30:32

Find an online example of a compelling information visualization. Describe:

1) what kind of data tables the visualization is based on, including variable types 2) what kind of visual structures are used 3) why the chosen visualization is effective.

Include a URL to the visualization.

URL: books_everyone_1.png


Raw data table:

1) the visualization is based on nominal variables and a functional table

2) The visualization style is called a word cloud. Structures include book title size differentials, object color differentials

3) visual size of the book titles is especially used to represent corresponding nominal frequency values. The signification of title size is lost in the scope of a single object, but is only pognient in juxtoposition with other variously sized book titles. It is possible that the spacing of the words communicates a message, but of what it is not obvious.

Eric Wishart - 4/4/2013 21:54:08

1) The tables data is the categorical summation of screen time of a football game, averaged over four games. Independent variable is time. The dependent variable is type of footage being shown during the broadcast, given the time.

2) The image uses a 10 by 10 grid, which resembles a tv. Colored images fill each of the square representing a portion of the broadcast. A percentage text is linked to one of the colors to let the viewer know exactly how much of the broadcast the color is representing.

3) The chosen visualization is effective because it quickly draws your attention to the main point of the images, which is that broadcast football games shots are mostly that of players standing around. In green you can see that the actual game is only a small fraction of the time that you spend watching a game.

Aarthi Ravi - 4/4/2013 22:55:50

Geni is a good example of information visualization. It is a website that helps you build a family tree by either creating a new one or attaching yourself to an existing tree. The Data Tables used basically describe the family hierarchy and links between members in the hierarchy. For example, for a given member the variables would be Links to Parents, Brothers, Sisters, Spouse, Sons and Daughters.The variable types included are: Nominal: Name of member, names of the links- father, mother etc, Ordinal: Is married/Not Married and Quantitative: Number of children. The Visual Structure mainly makes use of the spatial property and uses the space to represent people in the hierarchy. It makes use of a Composition structure where the axes are aligned. The y axis defines your depth in the family tree. And the X Axis consists of all members belonging to the same generation as you.The visual structure is encoded with folding(the axes are continued as you drag the tree) and recursion(zoom in out is allowed) as well.It makes use of 2D Area Marks to represent a member in the family hierarchy.The visual structure represents all the members of the family tree in terms of a graph where every member is linked to the next, current and previous generations through lines.The visual structure also makes use of some graphical properties like color to differentiate between male and female. The chosen visual structure is effective as it makes use of all the three basic kinds of visual structures that is Space, marks and graphical properties. URL: need to login to see the tree)

Tenzin Nyima - 4/4/2013 23:04:06

One of the most compelling information visualization I have experienced in the recent past was the online Results of 2012 Presidential Election on I closely followed the election by visiting their website. It was quite mesmerizing to see that didn't even use fancy maps for visual structure to make the US map this compelling. They simply used two-colored one-dimensional map (blue for Obama and red for Romney). They also used a horizontal bar above the map that turns red and blue (color getting expanded from each side). The main data type they used was the Blue and red color (on the one-dimensional map) and the variable was the number of votes. Depending on number of votes, each state on the map was either red or blue. The visualization was effective, especially if you wanted to check which state voted for which candidate. And to see which candidate was actually winning, you can simply look at the horizontal bar on the top of the map.

Winston Hsu - 4/4/2013 23:11:59

I'm using codeinfographic2-21.png

(the two piecharts in particular)

1.) This visualization probably comes from data tables with a nominal variable of job type, and a number of students in each job type 2.) The basic piecharts map area to quantitative numbers in each job type. 3.) This visualization is effective because it contradicts what people would expect. The proportion of people studying a subject should roughly equal the demand for each type of knowledge. Its a simple visualization that most people understand well, and the stark contrast between the piecharts catches attention.

Tenzin Nyima - 4/4/2013 23:37:24

One of the most compelling information visualization I have experienced in the recent past was the online Results of 2012 Presidential Election on I closely followed the election by visiting their website. It was quite mesmerizing to see that didn't even use fancy maps for visual structure to make the US map this compelling. They simply used two-colored one-dimensional map (blue for Obama and red for Romney). They also used a horizontal bar above the map that turns red and blue (color getting expanded from each side). The main data type they used was the Blue and red color (on the one-dimensional map) and the variable was the number of votes. Depending on number of votes, each state on the map was either red or blue. The visualization was effective, especially if you wanted to check which state voted for which candidate. Without any effort to read, listen or count the votes, it is visually very effective to check who is winning in different states. And to see which candidate was actually leading in the presidential race, you can simply look at the horizontal bar on the top of the map.


Yuliang Guan - 4/5/2013 0:01:42

The information visualization example I chose is called “Starbucks and Mconald’s Infographic.” The link is as follows:

(1) Data tables include relations (cases by variables) and metadata. This visualization is based on both of them. All three variable types are included. For instance, in the left part, the sources of coffee beans, paper for cups, and sugar are nominal variable N (an unordered set); the two coordinate graphs at the bottom are two ordered sets which we call Ordinal variable O ; and the quantitative variable Q (a numeric range) in this visualization is the Starbucks stores and McDonald’s restaurants by country. In addition, metadata is also used in this example. On the side of each set of numbers, there is a short descriptive information about the meaning of data.

(2) Visual structures include spatial substrates, marks, and graphical properties. In my opinion, the following visual structures are used in this example: Marks: This infographic uses three types of marks, which are points, lines, and areas. Spatial substrates: The use of space is reasonable and efficient. A map on the top and a graph at the bottom, very clear. Distortion: This example uses a graph showing the stores that Starbucks owns worldwide year by year, and also shows the number of stores in each country. Therefore, both overview and details are combined together.

(3) The chosen visualization is effective since the designer reasonably makes use of data tables and visual structures. Consequently, the information is clearly showed to people. The two world maps on the top make countries visible so that we can clearly figure out how many stores in each country. Meanwhile, marks and related descriptions are used on the side of maps to let people understand the meaning of each mark. Besides details, the designer also give us an overview at the bottom of this example which makes it easier for people to have a rough idea. In a word, data and information transformation is very successful in this visualization.

Eric Xiao - 4/5/2013 0:03:28

Data tables based on my connections (names = nominal, number of connections = quantitative), how they're connected to me (by company, school, etc., nominal) and how many shared connections I have with each person as well as how many shared interactions we have.

The visualization is a spread out graph with a lot of nodes, the center node being me. This chosen visualization was effective because it demonstrates how my networks are connected together and groups them together based on proximity, showing not only how I'm connected to these groups, but also how these groups are connected to each other. They are very easy to distinguish based on color.

Erika Delk - 4/5/2013 0:28:48

1. The visualization at the URL above describes the earthquakes along the California-Nevada fault line. The quakes are shown on a map and colored according to the how long ago they occurred, and sized according to magnitude. I suppose the "tables" it is based on would be earthquakes in this region by location, by time, and by magnitude. You could argue that location is the independent variable and that the rest are all dependent variables. All of the variables are quantitative variables.

2. With regards to visual structures, the image makes use of a map, color coding, and size coding.

3. This visual representation is effective because it allows the view to see the relationships between multiple pieces of data at once with relative ease. While the same information could be conveyed with a series of charts, this representation allows all variables to be represented at the same time. Particularly, by using a map to pinpoint quake locations, we forgo the potential confusion a user might have if they didn't know where a particular place was.

Kevin Liang - 4/5/2013 0:51:17


1) There is not particularly "data" that is represented but mainly a hierarchy of how internet sharing works. One variable is the category type. The hierarchy contains different types of internet users, each participating in a different role. 2) A pyramid is used describing how something so small can be expanded so big simply by sharing. 3) It is effective because the shape of a pyramid has the smallest on top, and the largest on the bottom. We the people are the ones who share data the most and hence is why we are at the bottom. Things nowadays seem to become popular through word of mouth that somehow becomes a viral hit. This pyramid correctly displays how each role is important for the role on the bottom of them in the pyramid to continue the chain reaction of internet piracy.

Timothy Wu - 4/5/2013 1:10:41

URL for visualization:

I chose the "Islands of High Income" visualization from the US Census website. 1) One data table that the visualization is based on is a data table that associates counties or city areas with the median income in that county or city. The county or city area variable is a nominal variable because it denotes names of areas that are either equal or not equal to the name of that area. The median income variable in the data table is a quantitative variable because it is a numeric range, in this case a range from [$18000, $112000]. The rows would be the county or city areas and the columns would be the the median income.

There would also be a need for an additional data table that maps the name of the county or city area with an encoding of the geographical boundary of that county or city area. The city or county name would be a nominal variable. The encoding of the geographical boundary would be a quantitative variable because it lies in some kind of range of coordinates that encompasses the United States. The rows would be the city or county name and the columns would be the encoding of the geographical boundary.

2) The kind of visual structure used is a map of the United States with boundaries for the state lines. The visualization is interactive, so you can move a slider at the bottom of the screen to manipulate the bounds of the median income. Areas denoted by a certain county or city area that fall into the income criteria specified by the slider are colored in green. The green represents whether the median income is greater than the number selected on the slider. This green color is contrasted with the gray color of the United States map.

3) The chosen visualization is effective because there is a strong contrast between the gray color of the United States map and the highlighted green color denoting the median household incomes of counties. The state lines on the US map give the viewer a sense of within which state the highlighted green counties are located. In addition, the interactivity of the visualization with the slider lends to a heightened level of understanding because the user can manipulate the variable and see the changes manifest themselves on the map in real time. The user can see the green pockets appearing where counties are not as affluent as you move the slider down towards median income $20000. Conversely, you can see many of the green highlighted pockets on the map disappearing as you move the slider closer towards median income $100000.

Monica To - 4/5/2013 1:14:20

URL to visualization:

1). The visualization I chose is an interactive flash infographic detailing the occupational outlook in America to see and visualize the effects of education level and careers and opportunities. The data tables shown in first screen of the the inforgraphic is a stylized bar-styled graphs where the x-axis has a row of triangles that represent different education levels and the y-axis or height/size of the triangles represent the amount of lifetime income people in these different education levels make. The education level is a nominal variable type and the lifetime income is a quantitative variable type. This data table takes the raw data and maps and groups people and their incomes into similar categories to show a correlation between the two variables. The infographic is also interactive so you could view more active diagrams if you click on "explore job openings by occupation". There are clickable buttons on the left that allow the user to filter the information they want to see.

2). The visual structures used in this visualization have a very contemporary and stylized aesthetic. The first diagram displays the amount of lifetime income earned by people who achieve a specific level of education. This diagram is a bar-styled graph where the color helps users differentiate the different education levels and size relative to each of the other triangles represent the amount of lifetime income earned. Going to the next diagram by clicking on the "explore job openings by occupation" button, you could now click an occupation category to filter the information. After a specific occupation is selected, the diagrams to the right in the form of a stylized pie-chart change to reflect the selected occupation category. The diagram has color that coordinates with the variables, which in this case is the education level. The contemporary pie-styled chart then displays a numerical percentage that represents the portion of job openings available to the different education levels. The visual structure is chosen because it shows triangles of the same size and for each education level the triangles fill up (similar to pie-charts) depending on the coordinating percentage.

3). This visualization is effective because it contains active diagrams. These allow the users to toggle back and forth between certain conditions or variable types to better understand the data. In the second diagram with the pie-styled charts. Being able to see the triangles "filling up" as the percentage increases and "filling down" as the percentage decreases allows the user to better sense the overall importance of the percentages and how certain percentages correlate or are affected by certain conditions. Instead of reading a list of percentages of job openings and their linked occupations, a user could glance at the visualization and take in the whole picture. With the use of color, relative sizing, dynamic changes, toggleable buttons, and easy-to-read numbers, a user does not need to rely on memory or see the difference between two different variables. As noted in the reading, the visualization aids in reducing the search for information, provides users with processing resources, and uses visual representations to enhance detection of patterns.

Edward Shi - 4/5/2013 1:18:30

The data table has one input variable mapped to two output variable. It takes the leaf elevation and maps it to assimilation rate. The elevation rate is a Quantitative Variable as are the assimilation rates. The visual structure used is a line graph. In the visual structure there are marks (lines), and has spatial markings based on phyotmeter number and light absorbed. The visualization is effective as it shows the trend of the effect leaf elevation very clearly. With numbers it is hard to get a general sense but seeing the line rise and fall is easy to tell the trend of where the elevation is optimal and where it is not sufficient.You can also tell if there are differences based on the seasons as you simply need to compare slope. It is easier to tell if one line looks steeper vs calculating slope for every single number.

Brett Johnson - 4/5/2013 2:01:10

How Far is it to Mars? -

1) This visualization is based on distances from the earth to the moon and Mars.The variables types used are distances in meters as well as pixels, both for the distances between the planets/moon as well as the planet/moon size. So, Mars is 53 pixels wide(over 6000 m).

2) No graphs are used, only the planets and moon separated by the correct distances. The interesting aspect of how this distance is interpreted is in the motion that the visualization uses. From earth to moon, the screen scrolls down in a relatively short amount of time. But for Mars, it takes an extremely long time.

3) This visualization is effective because besides being aesthetically pleasing, the different amounts of time spent scrolling serve to emphasizes the extreme distance. Because the data is so limited, there is almost no searching that the user has to do, and not really any patterns present. What the author set out to do, show the relative distances in a novel way, is done well by keeping the design simple.

Yong Hoon Lee - 4/5/2013 2:03:46

The visualization that I chose is from a website called, in the article found here:

The visualization is the first graph on the page, describing Yu Darvish's four-seam fastball averages and ranges over the course of last season and this one.

1. The data tables that this visualization is based on consists of pitch data for a pitcher named Yu Darvish, playing for the Major League Texas Rangers. The tables describe Darvish's four-seam fastballs throughout the 2012 season as well as from a game played on April 2, 2013. The variables are time, on the one hand, namely when the game was played in which the given pitch was thrown, and velocity, namely how many miles per hour the pitch was clocked at. In other words, this table has only one row, namely velocity, and many different cases, for each game that Darvish threw. Clearly, velocity is a quantitative variable, being a measure of some value, while time is an ordinal value in this case, as the times are all discrete days interspersed throughout a year. Indeed, what the visualization is concerned with is not necessarily averaging times of games, or anything of the sort. Rather, it is concerned with charting his velocities over time, a characteristic of an ordinal variable, in which what matters is the order of the data points, not necessarily their exact values.

2. This visualization uses the visual structures of points and lines plotted on an ordinal axis (the x-axis), and a quantitative one (the y-axis). In particular, the lines indicate the range of velocities thrown during a particular outing, and the green dots indicate the average velocity. The dots are colored differently than the lines onto which they are superimposed in order to increase their contrast, and the ordinal axis is split in two, with 2012 having a different shade than 2013, so as to visually split the time periods. In addition, there are guidelines in the quantitative axis which serve as estimators for the eye, so that it can see how close a certain point is to a given velocity easily.

3. This visualization is effective because it conveys two somewhat disparate points in a very elegant, minimalist presentation which is very easy to follow. First, the lines clearly indicate ranges, and one can easily tell from a glance that the range for the start in 2013 is greater than most of the ranges from 2012, for instance. The different coloring of the different years is very effective in splitting the visualized data points, and while they are labeled as such, the colors seem more natural than the actual labels in splitting the data. Furthermore, the green dots also clearly indicate averages, so that a higher green dot implies a faster overall velocity. In this way, the visualization combines two different dimensions, namely range and frequency, in one image, without burdening it with extraneous details such as in a box-and-whisker plot. Indeed, this is very similar to a box and whisker plot, but by stripping down the visualization, the creator enhances its message and makes it clearer.

Lishan Zhang - 4/5/2013 2:14:36

1) what kind of data tables the visualization is based on, including variable types

The data table contains the areas in San Francisco and the corresponding quantities of various crime types such as larceny, narcotics, assault, vandalism, warrants, prostitution, vehicle theft and robbery.

The variable types are:

  • Crime types (Nominal)
  • Quantities of crime rates in different areas (Quantitative or Ordinal)

2) what kind of visual structures are used

  • Spatial Substrate: The visualization maps the quantities of crime rates in different areas into the map of San Francisco as the most dominant dimension.
  • Marks: Using Areas and Volumes to mark the quantities of crime rates.

3) why the chosen visualization is effective.

  • Fast to interpret: The visualization uses height to represent the quantity of the crime statistics, which is quite obvious and easy to understand by people.
  • Fewer errors: People usually only care about the general amount rather than the specific number. The visualization will give people correct understanding of different crime types happening in San Francisco.
  • Conveys more distinctions: People can easily compare the difference situations of the crime types from the elevation model in the maps so that they can know the severity of crime types in different areas.

Moshe Leon - 4/5/2013 3:07:35

My link: <> My online example of compelling information visualization is the only image in this link. I think the Data Table the visualization is based on is a simple Variable --> Value relational table, such as the one on page 19, 1.4 or 1.5. The variables associated are: The expense name --> Nominal, the expense amount in dollars --> Quantitative, the expense size --> Ordinal. The visual structures used are circle surfaces (disks) in different sizes, which most resembles the graphical properties, and is utilizing another way of representing its content- a way which is refreshing and new. The chosen visualization is effective because the mapping of the data table is represented in a very expressive way, depicting anything we might want to see out of the information associated with this picture. I think that this information was in a different format, like a bar graph, with y-axis as the dollar amount, x-axis with the different expense names, and the size of the bars is then taken into a different visualization and transformed into different sizes disks. It is much more entertaining, and another element was added as smaller disks which I think represent all the smaller expenses before they are gathered to create the total one.

Soyeon Kim (Summer) - 4/5/2013 3:32:29

1) It would have data table of locations (longitude and altitude) of craigslist housing posts and list of the postings. Variable type for these numbers would be quantitative (more specifically float).

2) Location probes that use location in a visual structure to reveal additional data table information.

3) This type of visualization is very effective because when people are looking for housing, location is one of the big deciding factors. It is very inconvenient to see the listings first categorized by the city names (which is too broad) and then realize the exact location and then search for another. This visualization allows users to filter out the postings in a sight whether it is within their desired living location or not.

Tananun Songdechakraiwut - 4/5/2013 4:46:50

URL <>

1. The data table is also on the URL page. All variables are of type 'Quantitative'. 2. It is a diagram with x-axis as pipe dimension, and y-axis as heat loss. Each temperature is distinguished by a unique color. 3. It is effective because of its easily perceived and conceivable characteristics. In particular, the diagram clearly maps pipe dimension and its corresponding heat loss. Also, it separates each temperature by matching it to different color and put everything into one diagram. Thus, its complete picture as a whole(pattern when pipe dimension is increased implies higher heat loss and how temperature between pipes and surrounding air affects this pattern) can be seen and interpreted at ease.

Brian Wong - 4/5/2013 5:14:56

This visualization is from the New York Times and is called "Where the Heat and Thunder Hit Their Shots", displaying a color-coded, relative-sized, 2-d mapping of shots taken by the two NBA Finals teams last year.

1) The visualization is based on data tables that would record every single shot taken by every single player of both teams. The variables that would be recorded would include, the player, the x-y coordinate of the shot relative to a basketball court, the number of points of that shot, and if the shot was made or not. Values that could be gathered from this table include total shots taken by an entire team or a specific player, the number of shots in a certain location, the percentage of shots made in a specific location, and the average number of points scored in a spot, amongst others.

2) The visualization uses structures such as white space (to represent null or 0 values), color-codings from green to red to represent the success rate of a shot, sizes ('areas' in the reading) of hexagon-shaped markers ('points' in the reading) placed on a 2-d plane to represent locations and number of attempted shots.

3) The visualization is effective because it is placed on a 2-d plane that corresponds to the 2-d plane often viewed by someone watching a basketball game. Therefore, it is easy to associate "hot" areas on a court where a team makes a lot of baskets, compared with "cold" areas on a court. And the difference between these hot/cold success rate is effectively represented by the heat-color metaphor of hot is red and cold is blue.

Anh Mai - 4/5/2013 8:39:56

1) what kind of data tables the visualization is based on, including variable types

This is Akamai's network traffic visualization for major geographic regions in the world. The kind of data tables being used here is a functional table where the inputs are the name of the regions and the amount of traffic in bytes. The output to the tables are the percentage and color code on the map. The nominal variables are the name of the regions. Quantitative variables are the traffic in bytes and and the ordinal variables are the percentage and color code on map.

2) what kind of visual structures are used

This is a large-scale data monitoring with information map. Information is monitored and process in real time and displayed on the map.

3) why the chosen visualization is effective.

Because information is displayed on top of a world map, which already makes use of people's familiarity of how the world looks like and where each region is. The different coloring also catches the attention of users, instantly allowing them to recognize where the most network traffic is coming from.

Zhaochen "JJ" Liu - 4/5/2013 9:38:14

Error creating thumbnail: Unable to save thumbnail to destination
Source: We Are Number One!


1. What kind of data tables the visualization is based on, including variable types?

The data table is a table that displays each country and their corresponding percentage (the percent of the population older than 15 with a body-mass index greater than 30). There are two types of variables:

  • Country (nominal)
  • Percentage of population with a >30 BMI (quantitative, continuous)

The data table is something like this:


Percentage of population with a >30 BMI









It is ordered by the 2nd column.

2. What kind of visual structures are used?


This visualization technique utilized perception very well. It used the body size of each comic character to represent the obesity rate in each country. The bigger the body is, the higher the obesity rate is. The bigger body directly means that the people in that country have larger sizes.

Spatial Substrate

The visualization lay out the obesity country by country, ordered by obesity rate. It gave the reader a very clear signal what are the top countries in terms of obesity.

It also uses some special marks (the flag of that country) to let people easily visualize what countries they represent. So, for someone who are not familiar with the English name of each country, they can read the flag if they are more familiar with the countries’ flag.

3. Why the chosen visualization is effective?

Perception: size of each body represents the obesity rate

Order: they are ordered by obesity rate, descending, so people can easily find out which countries have the highest obesity rate

Flags: flags help people read the visualization more quickly, even if they don’t understand the English word

Clear: no two elements are graphically connected, unlike a lot of other visualization. Each body is an individual identity.

kate gorman - 4/5/2013 10:01:20 1) The data tables the visualization are based on are % of total sales from 2010. The variable here is the type of cookie sold. 2) Visual structures here are actually the actual cookies, which are used to accurately depict a pie chart 3) The chosen visual is effective because it allows the viewer to quickly match the portion of the pie chart to the cookie type, without thinking too much about it. It's also just very clever and eye-catching.

Alysha Jivani - 4/5/2013 10:44:12

World Mapper ( is a resource that I learned about in a Development Studies class that I took at Cal. It creates distorted cartograms by manipulating the areas of the countries based on whatever variable is selected.

For example, I decided to look at the map on Absolute Poverty. According to World Mapper, “absolute poverty is defined as living on the equivalent of US$2 a day or less.” I’m using the poster version because it’s a bit easier to look at than the webpage version:

(1) This map is based on a table of countries (nominal category) and the proportion of people living on less than or equal to US$2 in purchasing parity power a day (quantitative). (Side-note: Purchasing Power Parity is a way of adjusting the exchange rate between countries based on how much money would be needed in each country to purchase the same goods/services).

There are also two other forms of data visualization on this page that help with interpreting the map: (1) the bar graph indicating countries (nominal) and the population in millions of people living in absolute poverty (quantitative) and (2) a table showing territories with the highest absolute poverty proportions (ranking (ordinal), territory (nominal), and proportion of people living in absolute poverty within the country/region (quantitative)).

(2) The cartogram (map) uses color and area to display the data. At first, I was a little unclear on the color scheme being used. I believe that World Mapper is supposed to use the same color scheme on all of its maps to help divide the world into 12 regions to make the map easier to read. However, at first, I thought that the colors might be indicating the proportion of absolute poverty within the country, which might have been somewhat useful but actually might have also been redundant information. The cartogram tries to keep the relative position of countries somewhat similar to a normal map, so that people can still interpret which country is which, and it uses the visual structure of size (for the area) to reflect the proportion of people living in absolute poverty. A region with a larger area means that it has a higher proportion of the world total for the variable being shown (e.g. absolute poverty).

(3) This cartogram is effective because, at first glance, it’s very easy for a person to see which regions have the highest proportion of the world total for absolute poverty. Very quickly, one can see the relative size of regions like North America and Southern Asia and see the disparity between them. I think having the supporting visualizations (even if their designs aren’t novel) present is also helpful because they give an idea of the countries with the highest rates of absolute poverty (which might not be reflected in the area of the country on the map if the population of that country is smaller) and the actual number of people living in absolute poverty in various regions (the bar graph). I think the bar graph is useful because proportions can seem a bit abstract and, by giving actual concrete numbers, it helps put the proportions into perspective.

Joyce Liu - 4/5/2013 11:03:16

A vegetable seasonality chart: Kitchen101-Seasonal-Vegetables1.jpg

1) The type of data table this visualization is based on is a a bar chart with nominal dependent variables: winter, spring, summer, and autumn, as each season is its own category. The independent variables, the vegetables themselves, is also nominal, in the sense that each vegetable is a category itself.

2) The visualization utilizes the Gestalt principle of continuity and proximity in that the time in which the vegetable is in season is characterized by a continuous arc, and the arcs that are close together represent vegetables that are in season at around the same time. There is also a nice use of color, as the color of the arc is matched with the name of the vegetable. What perhaps could have been more effective was if the different seasons had their unique color schemes—spring pastel-ey, summer bright, autumn warm, winter dark—so that the viewer is better able to differentiate between the different seasons, but perhaps that would introduce too much color into the visualization, and there might be some complications because some vegetables span several seasons.

3) The chosen visualization is effective because the viewer can just locate the quadrant of the current season and then trace upwards to find the vegetables that are in season. The color of the names of the vegetables also correspond to their actual color—broccoli, kale are in green, sweet potatoes in orange, etc. It also provides the months, which gives the reader more precision on when the vegetables are in season.

Claire Tuna - 4/5/2013 13:04:15

On my Wells Fargo account dashboard, one section is called the "Money Map". The "Money Map" page includes a section called "My Spending Report", which is a bar graph of month to month total spending. There is a horizontal line at y= the average monthly spending, and parts of the bars below that point are blue, whereas spending above that point is orange. The variable on the x axis is the month, which is nominal. The variable of interest, on the y axis, is the amount spent in $, which is continuous. The bar graph is an appropriately expressive choice because it implies moreness and lessness in the y values, which is present since they are continuous.

The structures used are bars, intersected by a line, either colored orange or blue. Because of perception and the tendency toward simplified shapes, we perceive each month as one rectangle, intersected by a line segment (the average), rather than two adjacent rectangles. This is the appropriate message, as everything in the rectangle is in the category of “spending during month X”, but a further division is made between “spending under average” and “spending over average”. One can quickly look at the graph to distinguish how much they went over average on a certain month based on how much of the block is orange. Orange/blue I think was an interesting color choice. Because orange is close to red, it is more closely associated to “stop”, “danger”, etc. I see how red would map onto “spending above average”, but orange is not an immediate choice. Blue, as the marker of “spending under average”, makes sense also because it is a color associated with calmness and serenity. However, I think green is a closer metaphor, especially when associated with red. Green and red map to go and stop, positive and negative. Blue and orange, albeit close to green and red, do not as succinctly express this relationship.

The current month’s spending so far is expressed in green. I suppose the difference in color is to imply that the data isn’t done/still in progress. I don’t know if I think this is completely necessary. I also don’t know what happens if your current month is above the average. Is it still green? Is it green and orange? I can’t find the answer to this. I think the current month’s spending would be fine matching the setup in the previous months, blue and orange, or they should all be changed to green and red, but the mixture of green with blue and orange, meant to signify the difference between the current month and past months, I think is unnecessary detail.

link : mmm_1spending.gif

John Sloan - 4/5/2013 15:58:04


1) This is a bar graph based on how many (in percent) people showed up to an election in Boston vs. how old they were. The y axis is a quantitative variable of percentage since it is ordered in increasing numbers. The x-axis is also a quantitative variable since it is ordered by age increasing by number (though it could also be argued as ordinal since they are group in age groups like an ordered set). The final variable is which election we are looking at data for. There are three different ones each in a different color. This is a nominal variable because it is not ordered.

2) The visual structure used is a histogram and is expressive because all of the data is used in the mapping. The histogram uses ranges of data that are ordered in increasing sequence. For each range there are three bars shown to distinguish between the three different elections, each in a different color.

3) It is effective because it is fast to interpret, shows distinctions (both between the different elections and across the age groups), and leads to few errors. It is very easy to notice that the percentage of voters per age group peaks in the middle, meaning much younger and much older people make it out to vote less often. It also very clearly shows the distinction between the public's priorities in elections. Most people will vote for the general election but few will vote in the primaries.