Ad data (beta): CRP unveils new tool for tracking FCC ad data

By: Robert Maguire

Today, CRP is unveiling a new tool that allows users to track political ad buys daily, as they are reported to the Federal Communications Commission.

CRP's new project is the latest of several worthy efforts that have sought to unlock this information -- crucial as it is to understanding how candidates, super PACs and dark money groups operate in elections. In 2012, Jacob Fenton (then of the Sunlight Foundation) created "Political Ad Sleuth," a pioneering achievement that became the go-to resource for researchers and journalists tracking ad buys reported to the FCC. Similarly in 2012, ProPublica's Free the Files project tracked FCC filings, and, like AdSleuth, it engaged users by asking them to help crowdsource the very messy data.

Prior to that, anyone wanting this kind of info had to go to individual television or radio stations to look at records of air time sold by the stations for political ads. That changed in 2012, when the FCC required the network TV affiliates in the nation's top 50 markets to put their filings in an online political file. The mandate expanded in 2014 to all broadcast television stations, and earlier this year, was broadened again to include cable, satellite and radio. But the data is unwieldy and disorganized -- effectively unusable for most journalists and other stakeholders without a technical background. The only alternatives available until now have been provided by private companies charging steep fees, effectively putting the data out of reach for most newsrooms, watchdogs and the public.

CRP aims to provide a free, open and much more usable resource, and this is our first step. Our database currently contains nearly 800,000 FCC filings going back to early August 2012. We have extracted text from PDFs for most of these filings and, in some cases will pull key text to track particular aspects of the data.

Sometimes these files disappear from the FCC's web site. So we are also saving a backup copy of every filing that comes in to ensure that the record -- and the content we've made machine-readable -- remains, long after the buyer requests its removal or the station itself deletes it after the allotted two years.

We are updating daily with new data; the number of filings we are processing every day has skyrocketed from hundreds to thousands since the FCC began requiring more outlets to file reports on political ads. To put that into context, there have been almost exactly as many filings uploaded to the database in 2016 as were uploaded in the preceding three-and-a-half years.

Because we are downloading and extracting text from thousands of records every day, many of the pages in our Ad Data (Beta) section will be fresh from one day to the next. In the coming weeks and months, we will continue to expand access to the data and make improvements, as resources permit. Users with questions, comments or custom data requests (including for full content searches of the parsed data) should contact us at, and we will do our best to help -- again, resources permitting.

Ad Data (Beta) 

Users who visit the new page will see the most recent 100 political ad filings received that day, with links to download the documents themselves. The stations listed in those top 100 most recent filings each link to a summary page of the last 50 filings submitted by that station.

At the top of the page, a box allows users to search using zip code or a two-letter state abbreviation. Due to the sheer volume of filings we are downloading every day, the results are currently limited to the 10 most recent filings from that state, but below that, we show tallies of the number of filings from each station in the last 60 days. Users can click any station listed and see the 50 most recent filings from that station.

For example, on Tuesday, a search of all stations in New Hampshire (by putting "NH" in the search box) reveals both "prebook" and a contract submitted WMUR-TV for a six-figure buy from liberal dark money group Majority Forward, which is linked to the Democratic Senate leadership. The group, which is supporting Democratic candidates in tight Senate races, has not reported any spending to the FEC in NH, but this data gives a glimpse into what the group plans to do over the next few weeks.

A number of dark money groups active on the left and right show up in the results of these individual station searches from day to day. In many cases, these groups, like Majority Forward, have not reported their spending to the FEC because they are calling their stealth political ads "issue ads"; those have to be reported to the FEC only during certain pre-primary and pre-general election windows. Using the FEC's reporting windows to engage in political activity under the radar is common practice among dark money groups, but with this new we aim to make the practice easier to track.

Stay Tuned

In the coming weeks, we will be adding new search and download functionality -- including options that allow users to search by federal and state congressional districts.

In the longer term, our goal is to create tools that allow users to view up-to-date, standardized data about politically active groups, to search the content of the filings for details like individuals or vendors involved with the groups, and to view FCC data linked to relevant FEC and IRS data -- particularly where it will help fill in reporting gaps for dark money groups.

We are now seeing more than 30,000 filings a week showing requests for airtime, contracts and invoices for ad buys, and other correspondence between stations and representatives of the candidates and groups. We will be making as much of that available in the coming weeks and months as time and resources allow for users who are eager to track ad buys in the final months of the 2016 election -- but who don't have the funds to pay a private media tracking service.

The main obstacle to getting this data up onto is the lack of resources for dedicated staff needed for processing and quality control. While CRP has made considerable progress in standardizing the data, it will be extremely labor-intensive to keep the data up to snuff on the site. To provide some perspective, CRP has a dedicated researcher who works on outside spending data from the FEC. This data is very complex, and regularly requires extensive troubleshooting and diagnostics -- all of which is aided by the fact that the FEC has standardized forms with machine readable data, assigned unique IDs for all political groups, and codes for transactions in the data.

The FCC data, on the other hand, has almost no standardized forms, very little machine-readable data, and no unique identifiers for groups across stations. It's simply a much bigger task. CRP is committed to get as much of this value-added data as possible into the hands of reporters and others who need it -- and as quickly as possible. Meanwhile, we hope you'll dive into the data, let us know how you're using it and stay tuned for these and other developments yet to come.