Move UP wants to get involved in data visualization because they believe that it will be an effective way to enhance adult literacy programs in the Hartford Area. Their goal is essentially to help visualize how the demographics pair up with the information about the individuals from the programs in order to see how you can most effectively enhance and promote the programs.
Using the one set of data that I have been given I have been able to experiment with several different types of visualizations and learned how to use multiple different sources in order to assist me in creating the visualizations I want. Most of the visualizations that I have created so far this semester have been maps, or are going to become maps, but I have also made a pie graph (shown below). Depending on what future data comes my way, I could potentially look into making more visualizations beyond just the maps. In terms of the mapping, I have been able to get some basic maps, both layered and not (layered shown below). I am currently working on more layered maps that get the demographics involved, which is getting us much closer to our final goals.
Visualization #1:
Visualization #2: http://slkeane.github.io/Fusion-Layers-Wizard-1/
update by Jack on March 18th: this is an embedded iframe of your dataviz above
Sarah K – Thanks for your progress memo. You have successfully used the Google Fusion Tables layer wizard to integrate points and polygons, and I have suggestions on next steps.
First, you and your partner need to read and discuss our privacy guidelines at http://commons.trincoll.edu/dataviz/public-private/. If the point data layer in the map above is actual literacy client data (not fictitious), then it shows the street address, age, and literacy program of individual clients. While names are not displayed, it would be relatively easy to identify individuals with this data, which is probably not the goal of the program.
Possible solutions:
1) Password-protect this WordPress post while you’re working on it and share our seminar password with your partner organization. Also, make sure the individual point data Google Fusion Table sharing setting is “unlisted,” or even better, private or deleted (which removes it from public view altogether).
2) Discuss with your partner a better way to publicly display client data, such as:
a) geocoding individuals by zip code boundary (which is easy & private, but less precise)
b) geocoding individuals by census boundary (such as tract or block group; more precise, but requires more tools)
In either case, you can aggregate (or group together) individuals to report the number of clients per boundary area (or better, the percentage of clients among the total population of the census boundary, using your Social Explorer data). If so, this would transform your data from points to polygons, similar to the “Open Choice” map by Savvy Reuben on this page http://epress.trincoll.edu/dataviz/chapter/why-writing-about-data-matters/
Sidenote: if you zoom out, you’ll notice that one of your CT data points has been accidentally geolocated inside NJ, because it could not definitively place “Bloomfield” without the proper state. Be sure to create a “state” column or restrict your geocoding to CT in GFT, and redo.
Sarah has been doing an excellent job working with the data that I have been able to get for her. She has experimented with different maps using the data I’ve given to her, as well as found maps with the demographic overlays that are so important to making the connection between poverty and literacy rates. I also appreciate her ability to work independently and be flexible with our meeting schedule.
The next time we meet, we can talk more about the privacy issue. I think the best direction, (as Jack suggested), is to aggragate the data to the census tract or neighborhood level in order to maintain privacy of the addresses.
I will be attending the event on April 21st from noon -1.