Projects

 
 
 
 
 
 



Projects related to visualizing geographic data.



"Where are all of the trees?"
Here they are.

Have you ever wondered where all of the trees in San Francisco are located? I haven't, but this visualization is interesting nonetheless - see all of the trees within a certain radius of some location (very few in the Sunset, lots in the Fillmore), or identify patterns based on species (lots of plums on Divisadero).

This is an class project from about a year ago, explaining why it contains a lot of features and little analytical purpose.

Check out the interactive visualization here.




"Trees, Bikes, and Crime"
Stanford, by the data

In 2012, Stamen Design published a "Trees, Cabs, and Crime" map of San Francisco. The map overlaid the locations of trees, cab routes, and reported crimes in an attempt to present an image of the city through a unique lens. The visualization displayed the social and atmospheric geography of San Francisco, allowing its viewers to see the shapes of urban and natural spaces, while keeping the map uncluttered by removing street names, neighborhood divisions, artificial borders, as well as the various associations viewers have with those constructs. I decided to do something similar for Stanford.

While official maps of the campus produced by the university are often overloaded with street names, bus stops, and building numbers, such maps convey very little about the character and atmosphere of the places they depict. By contrast, I hope to help viewers discern the characteristics of different locales in Stanford’s campus, from the bustling quad to the large plots of undeveloped wooded land. Although the different environments may seem obvious to the students, faculty, and others who spend their days on the campus, my map formalizes these environments and grounds them in data.

The data-richness of the map means it takes a while to load, so here is an image of it. If you desire the original in all of its SVG glory, it can be found here. Enter at your own risk. I've had to restart my computer more times than I can count for that page. To strike a balance between quality and mental stability, check out this pdf version.




How could we change American values...
without changing the people?

Easy - rearrange the state boundaries. Right now, many of America's states are diverse, containing many groups of people with different values and ideas. This makes them a relatively successful method of governance - large majorities are less likely to dominate, and state governments can look out for minority interests better than the federal government can, simply because there are fewer small minorities in each state than in the entire nation (obviously)

Here is a set of maps with the United States divided into 53 (annoying number, I know) states divided up in just the opposite way. Instead of diverse states, I've divided the USA into groups of similar people. A nation divided as such would fall much more easily into just that - a nation divided. Below, the new states with some identifying cities, and the new states broken up into regions:



I also thought it would be interesting to see how the 2012 election played out in Alternative America. Turns out it's very similar - it would be more interesting to see the makeup of the Senate and the House under this system. And yes, there are fractional delegates. No, this is not real so no, I do not care.



"Would Trump still win?"
A return to Alternative America.

Now that Trump has won and we're living in Alternative Facts America, I've returned to my own version of Alternative America (outlined below) and created a new map to measure the results of the most recent election.

Turns out Trump still wins. He takes 57.7% of the electoral college vote (pretty similar to the real thing, actually, where he took 56.8%).





"Where is the music?"
The locations of Top 10 artists around the globe.

The recent "Where is the Art?" project (see the art page) turned out so well that there is now an accompanying project for music. However, this project is not only a modification of the previous one, but an evolution. Instead of mapping historical musical pieces (how do you figure out which are the best? or most famous? etc), this project maps the origins of the Apple Music-featured musicians and bands. The reasons it uses Apple Music (and not the Billboard Top 10 or something), are a) that Apple Music changes more frequently, so the project wouldn't repeat artists/songs nearly as often; and b) that I had Apple Music open when I conceived the idea for the project.

Another major difference is the way the map was made. The previous map had little squares that stacked kind of neatly into little towers of varying sizes. While it was kind of cool-looking, it was not very precise and didn't scale very well. This time around, I went with the data program Palladio, which is an interesting data-visualization software with a nice geographical feature. While this took a lot of extra data-wrangling, I think it was worth it.

Palladio lets you display different subsets of your data, and it works great for non-massive, geographical data sets (just like mine!). Combining a few weeks of data (colored by musical category) looks like this:



Crikey, that's a lot of stuff going on. But never fear - filterman is here. It simplifies if you look just at the US, and take away the lines that connect the genres, as you can see in the second map.

It's colored by genre:
  • Red - Overall Top 10
  • Pink - Latin
  • Gray - Hip Hop/Rap
  • Teal - Metal*
  • Green - Electronic**
  • Orange - Country
  • Yellow - Christian/Gospel
  • Blue - Blues
* Most top metal artists are international (see: Sweden), explaining the genre's small presence on this map.
** Just like metal, electronic artists are predominantly international, hailing mostly from Britain.

However, the real cool part of this project is that you can play with it yourself. Not only is the data in excel format here (bonus - extra sheets with different data arrangements, yay!), but you can take this data and play with it in Palladio. Go to Palladio, click the "Load Existing Project" tab, and load this file.