Select Page

DSSG: Vacants, Bikes, and Hairballs

By Rob Mitchum // January 28, 2014

Less than one week remains to apply to be a fellow or a mentor for the 2014 Eric & Wendy Schmidt Data Science for Social Good Summer Fellowship. In recent weeks, the DSSG blog has posted multiple recaps of the projects executed by its Class of 2013, working with government agencies, nonprofit organizations, and foundations on problems that really matter.

One team worked with the Cook County Land Bank Authorityto realize their mission of revitalizing neighborhoods plagued by vacant and abandoned properties. In their first post last summer, Juan-Pablo Velez described how the collapse of the real estate market and ensuing foreclosures created a crisis of vacant properties in the county, and how the Land Bank hopes to turn around this decline. A new post, written by fellow Skyler Whorton, presents the tool that the DSSG team created to help the Land Bank maximize their impact, highlighting those properties where reclamation or demolition will likely have the most benefits to the surrounding community.

With guidance from experts at DePaul University’s Institute for Housing Studies, we computed indicators for each community within the county, and then put them in context by ranking them relative to comparable areas. We also assembled some property-specific characteristics such as vacancy complaints, crimes, and recent sales and foreclosures. We used these data to model aspects of the housing market: by seeing how foreclosures spread, by finding the average price impact of amenities and disamenities, and so on.

The CCLBA Board of Directors can use these data and models to ask questions about each candidate property: Is this a parcel that is likely to sell on its own? How much demand for is there within the property’s sub-market for rental housing, owner-occupied housing, and retail? Does the area lack some amenity that a government agency (like a school or park district) might be able to supply? And how might treating this one property affect its neighbors?

Another new initiative in the Chicago area is Divvy, the city-wide bike share program launched last summer by the Chicago Department of Transportation. As a secondary project, six DSSG participants worked with CDOT to proactively address a nagging problem of all bike share programs: rebalancing. To make sure bike stations don’t get too full or empty, Divvy employs a fleet of trucks to move bikes from full stations to empty ones. DSSG fellow Adam Fishman writes about how the DSSG team used data from the Washington DC bike-share system to create an app that predicts when full or empty stations will happen in the future, allowing system dispatchers to deploy rebalancing efforts ahead of time.

Dispatchers are busy. They must simultaneously navigate the current state of the bike share system, the location of the trucks, city traffic and their own predictions of where the bikes are moving. While our models can help them with predictions, they don’t have the time to look through statistical output, p-values and confidence intervals. They need something simple that will help them do their job on the fly, something like a map of the predicted state of each station. Good news! That’s exactly what we made.

In order to succeed, nonprofits need to make the most of every opportunity available to them. But discovering and locating all potential partners and donors is no easy task. One 2013 DSSG team worked with the Case Foundation to realize their ambitious vision of The Giving Graph, a connected web of social organizations that helps people connect with the causes they care about. Fellow John Brock walks through how the team used Twitter, news articles, and mission statements from 2,000 different nonprofit organizations to create a “hairball” of connections to help nonprofits understand where they fit in the nonprofit ecosystem.

Now we can finally answer the question, “Which companies support nonprofits similar to me?”: By comparing mission statements, Twitter followers, and tweets, we found similar nonprofits. And by analyzing news articles and nonprofits’ webpages, we found companies that support those similar nonprofits. If you’re the executive director of Dogs Deserve Better, you can now see that 3M and Native Foods Cafe are potential supporters.

Of course, there are other questions that nonprofits care about. Other areas to explore include nonprofit collaboration (with whom should I write a grant?), branding (how is the language I use similar or different from the language used by other nonprofits?), and board member connections (to whom am I connected at other nonprofits or companies?). All of these questions are answerable from data that are already out there on the web. It’s just a matter of collecting them and finding the links between those pieces of data.

For more on DSSG projects, visit the fellowship blog or GitHub page. And if you would like to come to Chicago and join this summer’s fellowship, remember to apply by February 1st.