As any digital marketer knows, Facebook has a treasure drove of user data that makes it an immensely effective tool for serving micro-targeted advertising with precision.

The Facebook data team has leveraged this data in all kinds of interesting ways. Their annual mapping of the geographies for MLB and NFL franchise fandoms have gone viral (and revealed the true irrelevance of the New York Jets). But that’s just the tip of the iceberg — they’ve investigated the most pressing issues of our day in highlighting the age and gender differences in whether or not the dress was black and blue or white and gold.

The Facebook team has also looked at some very interesting political questions: which television programs over-index with partisan voting patterns and who is being reached by campaign’s organic social media posts. But what about using Facebook’s data to consider the political geography of the United States?

That’s why we’re here. Using the audience size estimate tool in Facebook’s Power Editor, it’s possible to determine how many individuals have liked any page related to the Democratic or Republican party on Facebook in each Congressional district.

Thanks to the invaluable effort of Rising Tide summer intern Leilani Harris, we can now present that data in what we’re calling The Facebook Midterms. Click here to view it in Google Doc form, with Daily Kos Elections’ 2012 presidential results by district presented alongside for comparison.

The Facebook Midterms

Image by Lucky Bommireddy, licensed under the Creative Commons Attribution-Share Alike 4.0

Dark red: Republican-held, no change | Dark blue: Democrat-held, no change | Light red: Republican takeover | Light blue: Democrat takeover
Click here to view in high-resolution.

In The Facebook Midterms, where Congressional representation is determined by Facebook like activity alone, 59 seats change hands, with Democrats netting a total of 3 seats (and another 24 contests being too close to call).

Democrats see their greatest gains come in large blue states: in The Facebook Midterms they would net 8 seats in Pennsylvania, 5 in California and New York, and another 4 in New Jersey. While Republicans would also consolidate support in states that traditionally lean their way — netting 6 seats in Texas and a number throughout smaller southern states — many of their victories would come in states that have recently been trending blue: Republicans would net 3 seats in each of Minnesota, Arizona, and Colorado.

While Facebook social data does not offer as complete a picture of the electorate as a traditional voter file or other more sophisticated data sources, it does offer insights into political geography that are not otherwise easily revealed.

Both the number of Democratic supporters (17.8 million) and Republican supporters (18.7 million) detected through this method are less than a third of the number that cast votes for those parties’ respective presidential candidates in 2012. It seems fair to conclude that these individuals, who have taken the affirmative step of liking a political page on Facebook, are more engaged and more politically active on the whole than their peers that have not.

As a result, this data may give us insight into where party activist communities have most taken root. In 2012, NY-15 and TX-13 were the most partisan Congressional districts, but per Facebook data, the districts with the highest proportion of party activists are Danny Davis’s Chicago-based IL-07 and Steven Palazzo’s gulf coast MS-04.

It would also appear as though The Facebook Midterms give us insight into the investment that state and local party committees have made to organizing on social media. Although LA-02 gave 75.8% of its vote to President Obama in 2012, Facebook data would forecast the district as a toss-up.

On its face, this is a particularly surprising result. However, once we dive in a little deeper an explanation presents itself: the Louisiana GOP has 70,000 Facebook likes compared to the Democratic state party’s 7,000. While the value of Facebook like investment is debatable, it is clear that some parties have prioritized it more than others. Because this investment will vary by district, it is difficult to know which factors have weighted most heavily in deciding any given seat.

All the same, that The Facebook Midterms remain so close in result to our actual midterms is just another testament to incredible user data that Facebook has and the importance it will continue to play in political advertising.