US Midterms, Twitter, Geotagging and what it may tell us

Yesterday saw the US midterm elections. As had been predicted the GOP picked up enough seats to take control of the Senate, while also retaining control of the House. As has become common, there were a lot of interesting graphics floating around. There were a couple I saw that together help to reinforce a point that I have been trying to make for a while.


Twitter have been doing a great job of drawing attention to interesting uses of Twitter by political candidates, and supporters. They have also produced a number of graphics around how voters are using Twitter. One that caught my eye was the following:


This graphic shows the distribution of geotagged tweets that contain #ivoted or I Voted. What I find interesting is the large geographic gaps. Especially when you compare this graphic to this one:


When you look at the states that had Senate elections yesterday vs the states that show up in the geotagged tweets there are a number of states that were missing. Add in the fact that all US states had House elections yesterday, and it becomes obvious that the Twitter is not representative of the whole country. There are large swathes of the country that have little to no representation in the geotagged tweets graphic. If Twitter is being used as a political tool, either to advance a message, or to try and measure the feelings of the wider community, this factor needs to be taken into account.


I haven’t come across much in the way of solid, in-depth analysis of the demographics of Twitter in New Zealand, but I suspect there are similar geographic areas that don’t have many users. I also suspect that there are demogrpahics that are over or under represented on Twitter. I would be surprised if most political parties haven’t realised this, but have they taken it into account when developing their social media campaigns?