Election Day Social Media Predictions

January 5 is finally here for Georgians eager to cast their votes on the US Senate runoff elections. The polls are all over the place, with most of them showing differences between the candidates that are within the margin of error. The site FiveThirtyEight published a final average of various polls which put Ossoff (D) ahead of Perdue (R) by 1.8%, and put Warnock (D) ahead of Loeffler (R) by 2.1%.

Continuing our previous analyses of the candidates’ social media accounts on December 21 and December 7, we will now present a final update of their social media data as of January 4 (eve of the election), and use the same methodology as before to estimate the percentage of likely votes to each candidate.

We used the Odysci Media Analyzer to monitor the social media pages on Facebook and Twitter of Jon Ossoff (D), David Perdue (R), Reverend Raphael Warnock (D) and Kelly Loeffler (R). We collected daily data for six weeks between November 24 and January 4.

Jumping to the Predictions

If you just want to see our social-media-based predictions, here they are.

Regular Election: Jon Ossoff (D) ahead of David Perdue (R) by 0.76%

Special Election: Raphael Warnock (D) ahead of Kelly Loeffler (R) by 1.58%

If you would like to know how we came up with these numbers, please read on.

Fans and Followers

As of January 4, 10am EST, the total number of fans and followers of each candidate on Facebook and Twitter, was as shown below. The relative positions remained the same as in our previous analysis on December 21 with Ossoff leading, followed by Warnock, then Loeffler, and lastly Perdue.

The figure below shows the number of fans and followers that each candidate gained in each of the past five weeks, and the percentage increase over the base number at the beginning of each week. The percentage gains of all candidates, last week (Dec/29-Jan/4), were similar (4.0% for Perdue, 4.1% for Ossoff, 4.7% for Warnock, and 4.9% for Loeffler), but the increases in absolute numbers were particularly impressive for Ossoff and Warnock, adding over 38K, and 29K fans and followers, respectively, in one week.

We also looked at the relative numbers of fans and followers gained by each candidate. For each 100 new people who started following any of the candidates on any given day, how many followed each candidate? The following figures look into these numbers pairwise for each senate race for the last six weeks.

The figure above breaks down the distribution of new fans and followers for both candidates, Jon Ossoff (D) and David Perdue (R), between November 24 and January 4, as well as the total in the period. Although there are some daily fluctuations, Ossoff is clearly maintaining a steady lead on the percentage of new fans gained daily. Overall, for the last 6 weeks (up to Jan/4), for each new 100 fans and followers, 83 sided with Jon Ossoff (D) and 17 sided with David Perdue (R).

The figure above shows the same metric for candidates Reverend Warnock (D) and Kelly Loeffler (R). In this race, the daily fluctuations were larger than in the regular race. Overall, for the last 6 weeks (up to Jan/4), for each new 100 fans and followers, 63 sided with Reverend Warnock (D) and 37 sided with Kelly Loeffler (R).

Using Social Media to Forecast the Election

In our previous blog we outlined the methodology used for estimating the number of votes to each candidate based on their social media data. For the sake of completeness, we will repeat the description here.

The estimation starts with the number of votes received by each candidate in the November 3 election, considering all candidates that ran for the Georgia US Senate seats (regular and special elections). Then using these numbers of votes, and the percentage numbers of new fans and followers that each candidate gathered in the last six weeks, we predict the likely percentage of votes that each candidate will receive on election day, January 5. Details of the methodology are as follows:

1. Get the numbers of votes received by all candidates in the first election (Nov/3). We will assume that all voters that cast a valid vote on November 3 will do so again on January 5.
2. We assume that the people who voted, in the first election, for Jon Ossoff, David Perdue, Reverend Warnock and Kelly Loeffler, will do so for the same candidate again in the runoff election.
3. In the regular election, besides Ossoff and Perdue there was only one other candidate, Shane Hazel, a LIbertarian, who got about 115 thousand votes. Since Hazel has not officially endorsed any candidate, we will consider these 115 thousand voters as undecided, but assuming that they will vote again on January 5.
4. In the special election, besides Warnock and Loeffler, there were 18 other candidates, from various parties. We will now “reassign” the votes for those 18 candidates to Warnock or Loeffler. We will assume that a high percentage (90%) of people who voted for a Democratic candidate will transfer their vote to Warnock, and the same percentage of people who voted for a Republican candidate will vote for Loeffler. The 10% remaining from each group of voters (pro Democratic or pro Republican) will be considered as undecided, or people who may switch camps. The people who voted for candidates in other parties (Libertarian, Green, Independent) will all be assumed to be undecided.
5. We assume that the number of undecided voters, in each election, will be the sum of the undecided voters as explained in steps 3. and 4. above, plus a number of new voters (who did not vote in the first round), which we will assume to be 75 thousand. This estimate is based on the number of new voter registrations between November 3 and December 7 (registration deadline).
6. Based on the steps above, we estimate the number of undecided voters in the regular runoff election as 190 thousand, whereas for the special election we estimate the number of undecided voters as 376 thousand.
7. We will distribute these undecided voters per day in the 6 weeks considered in this analysis (Nov/24 to Jan/4), as if they were making up their minds, a few each day; and we will “assign” them to each candidate (in each race, separately) based on the percentage of new fans and followers that they acquired each day, as shown in the graphs above entitled “For each new 100 new fans and followers, how many go to each candidate per day?“

The figures at the top of this blog illustrate the percentage of likely votes per candidate, per day, and the final value for the regular and special elections. The initial percentages on November 23 are based on the numbers of votes that each candidate received on November 3, plus voter migration as explained in step 3. and 4. above.

For the Ossoff vs. Perdue race, we estimate that Ossoff leads Perdue by 0.76 percentage points on January 4. For the Warnock vs. Loeffler race, our estimate shows Warnock leading Loeffler by 1.58 percentage points on January 4.

In both races the edge of one candidate over another is within the margin of error of election polls (usually 2% to 4%), so it can certainly go either way. It may be a few days or weeks before we know the final vote counts in Georgia, and it will certainly be interesting to find out how our simple social-media-based estimates fare against expensive and somewhat unreliable traditional polls.