Election DataBot: harnessing the power of the matrix

There’s no shortage of public data out there around the election — FEC filings, demographic data from the census, even search data from Google… the list is long and complicated. But imagine if, like Neo in The Matrix, you could plug yourself into that election feed and get a complete grasp of all the data in one place?

That’s where the Election DataBot comes in. We worked with ProPublica’s data team of Derek Willis, Ken Schwencke, Lena Groeger and Scott Klein to create a resource for key election data sources. Part of a bigger project called Electionland, the DataBot pulls them together into a single source combining key datasets for the Presidential election and congressional votes, including:

• Federal Election Commission filings: Every time money is donated to a candidate or spent for or against them, it’s recorded by the FEC, and the ProPublica Campaign Finance API pulls the latest filings into the Databot every 15 minutes.
• Google search data: Real-time Google search interest for each of the candidates in the 435 House races, 34 Senate seats up for grabs, and the Presidential race.
• Voting records and congressional data: Voting history for all incumbent candidates from the ProPublica Represent database of voting history and the ProPublica Congress API.
• Race ratings and forecasts: Which way is each seat and state leaning? Race ratings are from the Cook Political Report. Race forecasts are from FiveThirtyEight.
• Latest polling data: via the Huffington Post polling API, which collates hundreds of polls.
• Demographics: Data from the Census Bureau on the demographic breakdown of each area.
• Coming soon… Latest campaign YouTube videos added by each candidate, as they add them, plus API access.

How to use the Election DataBot

There are different ways to get the key data needed from the tool. In particular you can view:

The Firehose

Every time there’s a single new filing or data point, it shows up in the Firehose. This is the unfiltered equivalent of seeing the inner workings of an election happen in real time.

Trending candidates

Every single politician up for election is ranked to see whether they are trending over the past four hours — which means they have started accelerating quickly in search (you can read more about ‘trending’ data here).

Individual candidates

Every single candidate has a page of data associated with them. Each page has the latest data on that candidate, including Google search data, FEC filings and spending summaries. As of today, the page for the Presidential candidates tells you that of spending outside the campaign:

• While $19.1m had been spent in support of Hillary Clinton, $30.2m had been spent against her. See the Clinton page.
• $110.4m has been spent against Donald Trump — compared to $15.6m in support of him. See the Trump page.

Races

There’s a page for each state and national race where you can compare the candidates to one another — and see which way the polls are leaning — demographic data, Google Trends data comparing the candidates in search and more.

Notifications

Following a specific race? You can set alerts to notify you when a new filing or data point happens.

More filters

Click on a type of report and you can choose just to see them. So click on ‘polls’ and you get all the latest polls. Click on ‘trending’ and you can see all the latest trending Google search alerts.

Play with the data

The best data journalism is about sharing data and making it open and available, with that, all data journalism becomes stronger. In essence, it’s about creating a community around the data. And that community becomes even more important during an election campaign when the data is the stuff of democracy itself. That’s where the Databot comes in. And we can’t wait to see what you do with it.

Simon Rogers is data editor at the Google News Lab. You can read more about how to use the Election Databot here.