As encouraged, this is the simplest, most viable idea. Probably best if I’m not the strongest coder and running a cold for the week…

In the simplest form, I’d like my idea to visualize the growing or shrinking arguments on both sides (for or against the Chinese mother). The interface in mind is simple: like a video player. The user can drag between old and new and watch the blobs change.

When hovering over either blob at any given spot, the user will be given the most prevailing argument (as measured by the number of inline comments) and words (count frequency of word in the area maybe +/- 25 comments.

The blobs themselves will be driven by word counts as well… After figuring out which are the top 5 words in a 50 comment radius, the change in the count (positive or negative) will affect one point on the blob dragging it in forms. (Haven’t looked into what language to use, but *something’s* gotta be able to support this right?

This lava lamp-esque means of visualizing a firey argument also intrigues me.. If the blobs expand into each other, they will repel.

My idea for the project has evolved into two separate but related sub-projects.

My initial idea was to create a family tree with fun facts and juicy details about the lives of some royal family. It would include not only relationships but also highlight the more interesting parts of history including scandals, mental illness, and mysterious deaths.

Jumping off from suggestions in class that I look at families other than the royalty whose lineages have already been mapped out, I thought about comparing their family tree with that of some regular Joes. However, I don’t have easy access to my own family’s data (much less in a computer-readable form) and I would like to avoid using some random stranger’s family tree from the internet, so I’ve decided to construct a ‘typical family tree’ from US family statistics throughout the years. Specifically, I will be using data for life expectancy, household size, marriage trends, and most popular names. This could probably be enough of an information visualization on its own, so I might narrow down my scope to just this portion if I run short on time.

Questions to consider: How do I accurately convert statistics into one “average” family? Is there an easier way to get royal family relationships than going through Wikipedia’s articles? Will comparing the two trees actually lead to any insight, or should I just concentrate on constructing one or the other?

I am working with a log of all of my actions for the fall and spring of my sophomore year.

Below are some sketches that I made, beginning to think about the form of my visualization. They are mostly straightforward, looking at bar charts and pie charts. The bigger questions I’ve been asking is what I am comparing across — I think the interesting things will come out in looking at falls/spring or weekend/weekday, for example. I was also wrestling with the question of static/interactive. I’ve decided to work interactively, most likely in some kind of simple interface with a time line at the bottom, and the “visualization” at top. I am trying to figure out how to incorporate/build in small viz of a single day to the overall viz of the entire semester (so how does this specific day compare to the rest of the semester, or the “average” monday).

Next things next — typin’ up the data.

Tracking Taxi Flow Across the City

This is a really spectacular heatmap of Manhattan showing where taxis are picking up fares across the borough. An extra dimension of this visualization is the ability to watch the data change over the course of the week by playing the animation. Now that taxis in NYC have mandated GPS systems installed, very rich information can be extracted and mapped like this. Taxi pickups peak at major tourist shopping districts and transportation hubs such as Penn Station and Grand Central Terminal.

The United States of Autocomplete

This is a geographic map that attempts to show what people are thinking about when searching for state names in Google. The labels for the state names are all auto-completed to make interesting associations between the state names and popular searches. The map template used makes the labels difficult to read, but the version of “The People’s Republic of Autocomplete” addresses this problem and makes it easier to read.

Stephan Thiel made Understanding Shakespeare as his B.A. thesis project to provide a new analysis of Shakespeare’s works and invite his audience to re-examine their understanding of them. The visualizations were created using Processing and toxiclibs. They present the most frequently used words for each character broken down by scene. Characters are ordered by appearance in the scene from left to right. Presenting the characters of each scene in order of appearance preserves some of the natural flow of the story. The overall advantage to this format of presentation for text analysis of a narrative is clear, as it uses a reasonable chunk size (scenes) to preserve the distinct moments throughout the story. e.g.: With Romeo and Juliet, the balcony scene in act 2 scene 2 is full of love and warm-fuzzies, with vocabulary to match, but that is quite different from act 5 scene 3 where our two leads kill themselves.

[ Understanding Shakespeare ]

Poetry on the Road | Boris Müller

I like the abstraction of this piece. It is less of an “infographic”, finding a way to display the data in a visually appealing/informative manner, and more of a unique structure on its own, using the data merely as a means to an end. The work takes a poem and generates a single sheet of “paper” with multiple lines to make folds on. Each line represents a word that changes depending on its length and frequency. It is then folded in order to create unique forms, like a “mad oragami master”. As well as the abstraction element, I like the aspect that this piece would be so easy to bring into the physical realm. As well as a digital object, you could print out the flat “paper” state and actually create the maddening sculptures, maybe bringing them to an even architectural scale.

“View of the solar system showing the locations of all the asteroids starting in 1980. As asteroids are discovered they are added to the map and highlighted in white so that you can pick out the new ones.” Simple, clean and visually appealing. I especially like how you can pick out the differences between the technology they were using to discover asteroids in the different eras by the frequency and placement of the sparking new white asteroids.

The Stanford Dissertation Browser is an interface for viewing the relationships between Stanford departments based on a text model applied to dissertation abstracts. The Browser is an excellent model of presenting data flexibly and from multiple perspectives: Relationships are presented based on the selection of a department, and a different selection re-forms the presentation. It would take two times the number of departments times the number of years in simple pictures to present the same amount of data that this interactive visualization contains. This presentation method isn’t limited to dissertation abstracts or even text analysis, but could represent any relationship between nodes in a set over time or some other convenient form of progression.

[ The Stanford Dissertation Browser ]

Pa++ern – by daito

Datasets to use:
+ Comments on Youtube video or possibly 4Chan, parsing text to see what trends come from it
+ Facebook Network Packets w/ Carnivore PE or personal account data scraping w/ php
+ Game Engine Events
+ web stats – stat whore (combine networks, gmail, sms, analytics, YT hits, games, etc)

Questions to ask:
Does the data necessarily need to be visual – couldn’t it be audible, feel-able, smell-able?
My biggest problem with information visualization is the seemingly lack of human experience – the phenomenological.

Sorry if this offends anyone — but it asks critical questions of network culture

tits or gtfo from Chris Beckman on Vimeo.

So you think the internet is still free

This website briefly, but beautifully charts the different ways and reasons governments censor internet access around the world. Upon further inspection, the information it presents is varied and well scoped, but is not particularly deeply explored in each of the visualizations. The website also lacks proper citations (there are not even links to the sources it draws froms), which severely hinders both its usefulness and credibility on the internet. There are, nevertheless, several strong aspects of the visualization. First is that it works with different kinds of data – it shows different views and ranges of the information. Its visual language is very consistent, and works well with the language of the visualizations.

http://ngrams.googlelabs.com/

The google nGram viewer lets you quickly and easily see the occurrence of a word, or words, across time as found in google books. Its visualization, like that of Facebook’s lexicon, take the form of a graph, with time as the x axis, and word frequency as the y axis. The scope of the nGram viewer is what makes it interesting — seeing how a word falls in and out of relevancy. As a tool it is straight forward, with no frills. While this makes the nGram viewer easy to use, the tool could benefit from easy ways to integrate “social” media. Why is there not a built in community, for example? While fun to play with, I’m not sure as a viewer that I want to sit there and spend time looking for a particularly insightful pairing of words to look up, and I would enjoy the opportunity of having a community built in to google that let me view graphs that others found interesting.

http://www.informationisbeautiful.net/visualizations/google-ngram-experiments/

1. LDAP

Each student, staff member, and faculty member that comes to or works for CMU has some amount of personal info made available by default with the LDAP servers. The front-end of this service can be seen at cmu.edu/directory. The information includes a person’s associated departments, majors, student level, primary affiliation, name, andrew id, and email address. Analyzing this data might let us draw some conclusions about the probabilities of getting a minor with a certain major; the declaration of a minor by a certain student level in a particular department; or even the likelihood of someone in a certain major to forward their mail to Gmail.

2. Foursquare API

Foursquare is a framework and service for location-based social networking. The Foursquare API allows you to pull information about venues, users,  user checkins at a venue, and tips for a venue in addition to allowing you to perform actions like a checkin. Using that data, you could get an idea of where you and your friends typically go or what users who visit a particular venue are also likely to visit.

3. CMU Service Monitoring Data

Mail, OLR, Portal, LDAP, Calendar, Blackboard, KDC, WebISO, and smattering of web servers that keep this university running each fall down occasionally. When that happens, the monitoring system leaps into action and alerts the poor sucker on coverage duty to go and fix whatever’s broken. The service, problem, and alert timestamp are all logged, and visualizing that data over a variable time granularity might lead to some interesting observations.

1. Using Google Trends to self-explore my last years online searching behavior. From Google Trands I can get my searching keywords, searching frequency, and even searching time by months. I can visualize these data to analyze my online activities.

2.  Google Trends provide searching keywords datasets. Using this feature, I can explore people’s online searching keywords and do some interesting comparison, such as the trend of searching “Good” Vs searching “Evil”.

3. I am kinds of interested in the evolution of programming language. Since the very first programming language came out, many of them have died, some of them evolve to others, and some just become popular. I hope to visualize the trends of programming language’s evolution.

4. GLOBAL HUNGER INDEX DATA

It is good to know that on the same earth there are still that huge numbers of people who don’t have enough food and facing the threat of death. I found this dataset from here. Maybe I can do some comparison between the “hungry” world and the “wealth” world through some physical visualization installation, to arouse people’s care for the poor world.

1. Notabilia

Wikipedia has become a truly incredible compendium of free human knowledge, and a cultural barometer of sorts in terms of what we believe is worth knowing. Visualizing the discussions on removing items from its vast library is intriguing because the topics it shows aren’t always considered because they aren’t notable, but are sometimes considered because they’re things we’d like to culturally forget and ignore. The keep/delete leanings help frame and provide a story for each topic, but the fact that we see the story before the topic is almost a spoiler. I’d rather pick a name and then be surprised by what the community has voted for.

2. Tokyo: Right Now

Tokyo: Right Now is a visualization of census data collected in Japan of what people in Tokyo are doing. They asked citizens to record their actions in 15 minute intervals, and the data is available online to be browsed by gender, economic status, and activity over a weekly period in every minute. It’s amazing to see such a complete picture of a city in data, and to be able to browse it easily. That this data is so general yet can become so focused is very intriguing.

http://www.visualizing.org/visualizations/research-royal-college-art

I actually don’t think there’s much to be gleamed from this data visualization, but in a way it is a very very beautiful graphic. I guess the only thing that’s probably visible are which agencies are those that funds many of the phd students but the extent in which it differs from an agency to another is unclear. However, it is laid out in a pretty beautiful manner and the entire layout achieves an almost poetic feel to it. In short, I’m not sure what is the point of this visualization, but it’s just beautiful.

I’ve long wanted to record the 3-dimensional shape of a tree–both the high frequency movements caused by the wind as well as the slower growth. I’d be interested to see what parts of the tree move when. Does it swell uniformly, or does the top explode out while the base remains relatively static, or does the entire structure extrude out of the ground? Do the boughs mostly shoot outward, or do they writhe and warp over time? Also, applying that movement data to other 3D models could be interesting–imagine a teapot that flutters in the wind.

There are many apartment buildings around here. Each is a grid of windows, and in each cell, a domestic cycle plays out, indicated by the light turning on and off–a single bit of information. It might be interesting to point a camera at the façade and record the activity overnight. Each “Fourth Floor, Dawn, Up All Night Writing Letters” becomes a datapoint.

But I think that what I actually want to do is graph the color palette of a movie over its duration. I’m working on sampling pixels from a DVD rip and doing cluster analysis to construct a palette for each frame. Each frame’s palette is rendered as a stacked column of pixels. Columns are arranged horizontally to create an oscillating rainbow of color over time. Depending on how recognizable the palettes are, I could then make an interactive game of “match the movie to the palette”.

1. State of Balance: Artificial versus Natural

This project using a physical representation to exploring the complex relationship between the global emitted and absorbed carbon dioxide emissions.

See detail

I choose this project for several reasons:

1. among millions of graphic visualization projects, this one stands out for its special representation – using tangible objects to explore the data.

2. Good concept. By changing the rice in different icon objects, one can see directly from the balance changing that even a tiny little changing in the world can ruin the world’s balance. Very interesting concept and representation.

2. A Week in the Life



this is another physical visualization project. See details.

In this project, the data sculpture represents the author’s one week movement and communication made with his cell phone.

http://infosthetics.com/archives/2011/01/the_illusion_of_diversity_visualizing_ownership_in_the_soft_drink_industry.html

I’m not a big fan of the soft drink industry. In fact, I don’t really drink anything other than diet coke and I think soft drinks are really bad for human health. This data visualization kinda just highlights how absurd the entire situation is and how many sub brands are controlled by the big three. It’s just insane how big the entire sugared water industry is and how far they go to segment their product so that it reaches every part of the population.

1. Why Chinese Mothers Are Superior.
   This article set a cultural debate on fire, it’d be great to map the 7k comments or many rebuttals, somehow. Perhaps mapping some reoccurring ideas based on stereotypes.
2. Twitter as better than Jenny Craig.
   Use public tweets track what people have been eating, how much weight has been loss, how much support they’ve had … This is a prevailing hot topic of all times which will have a constant stream of new tweets.
   See related articles on ABC News and Lifehacker. It apparently works!
3. Catfight!
   I know cats on the internet are kind of overdone, but I still love it. Use Google Trends to provide mappings of popularity of certain memes, e.g. lolrus vs. happy cat.

I’m trying to think of creative applications of data visualization

I’m thinking of combining Twitter feeds and the Google Ngram datasets. One idea is to mine the latest tweets and see how often the significant words appear in Google books. How common words in tweets relate to book data and vice versa. Perhaps one cool thing that can be done with this is the kind of books that the tweets would recommend, so it can be a real time visualization where the tweet will break down to the common books that the reader would like to read.

I’m interested also in asking questions about education. I’m thinking if data.gov would have comprehensive data that can answer questions like how does education level correspond to GDP. or if student enrollment levels have to do with things like video game sales. Perhaps one big question we can ask is how crimes correspond to education and if that it fits the model that we seem to think it does.

The last thing I have in mind is slightly sillier. I’m thinking about doing a camera/kinect capture of the human silhouette and trying to plot a height and width graph of a person across the other. I’m just kinda thinking if there is something that can be done in this front and the actual visualization can use the actual silhouette to plot something interesting by itself.