My Favorite Things


Derek Willis


July 29, 2015

As first reported on the tweets, I have a new job, at ProPublica, where I’ll be working on news applications, investigations and other incredible stuff with the News Apps team.

Leaving The New York Times wasn’t easy. There are so many good people there, so many really smart journalists. So as a way of saying goodbye, here’s a list of some of my favorite things from my seven-and-a-half years at The Times (my longest time in any job, btw):

  1. Represent. Even though it didn’t last, this was the most rewarding thing I worked on in terms of coming up with an idea and seeing it to completion, thanks mostly to my then-colleague Andrei Scheinkman, now at FiveThirtyEight. Represent was a departure for The Times into personalization and feed aggregation at a time when both were still new ideas in the newsroom. It also was, to the best of my knowledge, the first Django app at the NYT. We stopped maintaining it a few years ago, but it lived on in other projects, including the Districts API and “The Other Races” from 2013.
  2. Toxic Waters. One of the first large projects I worked on at The Times, this involved both obtaining and loading EPA data as well as helping to decide how we’d present records of inspections and violations for water pollution permit holders. I didn’t do the really amazing visual aspects of this project, but really enjoyed digging into a new dataset and being able to help readers find out more information on water pollution where they live.
  3. Fech. This started as an intern project with Michael Strickland and Evan Carmi; the goal was to create an internal toolkit to process electronic campaign finance filings from the FEC. Since releasing it, we’ve gotten contributions from people outside The Times and Fech has become something of a standard, inspiring a Python version, too. It also powers Itemizer.
  4. 2012 Election Coverage and results. Elections are a huge deal at The Times, an all-hands-on-deck affair. I was really proud to be a part of the process for elections beginning in 2008, but for me 2012 was a high-water mark. We were able to build individual campaign finance elements like this one with Jeremy Ashkenas and another with Kevin Quealy, plus contribute to exit polls and other parts.
  5. Campaign Finance and Congress APIs. These were somewhat accidental projects in the sense that from a publishing standpoint they weren’t strictly necessary. But both APIs provided information to the public that otherwise would have been difficult to obtain. Even though the FEC now has an API and the United States project has expanded the congressional data available, they have been used in academic projects, silly projects and serious projects.
  6. Upshot Mississippi Senate primary coverage. The 2014 GOP primary and runoff in Mississippi gave me a chance to use precinct-level results data in near real-time. Pairing with Nate Cohn, we examined the impact of black turnout in two stories that provided evidence instead of just anecdotes. I also was able to delve into the murky world of outside spending in that race.
  7. VIGOP This is a crazy story, still ongoing, about how a small party organization is trying to raise its profile through some controversial and fairly unorthodox methods. The artwork makes it even better.
  8. Sunday Morning Talk Show Guests. One of my favorite things to do is to try and confirm or debunk the conventional wisdom about a topic, and there are few things more conventional wisdom than being a guest on the Sunday talk shows. I worked for several weeks standardizing data on talk show guests (which we released) and then helped build the table that displayed the most frequent guests.
  9. ActBlue. One of the biggest stories around political fundraising has been the rise of the small donor, and ActBlue, a Democratic conduit organization, is right in the middle of that. I used ActBlue filings to show how their users have expanded and grown as repeat donors, to the benefit of Democratic candidates and organizations up and down the ballot.
  10. Maiden Names. When I came to The Times, I half-jokingly said that one of my goals was to combine the wedding announcements with people mentioned in DealBook to create a network of New York’s elite. We never did that, but before I left I was able to finally do an analysis of wedding announcement data with my Upshot colleague Claire Cain Miller and get a Sunday Styles byline, to boot. This one was a lot of fun, and a great way to end my time at The Times.