Predicting the General Election

In the run up to the December General Election, perhaps more so than with any other previous election, tensions are high amongst parties and the public, as speculations arise regarding the country’s political preferences.

As ever, media outlets attempt to predict opinions and voting intentions, generating trend lines and averages to hypothesise the election outcome based on an average of the population. Despite polling companies claiming that 95% of the time, a poll of 1,000 people is accurate within a 3% margin, it was made clear in the 2017 election that this was not the case. Once again, the polls were not an accurate guide to the election result.

Look left, look right

The technology and innovation behind Text Mining Solutions, means it is able to reduce the margin of error, increase the level of analysis detail, and has the capacity to potentially predict the 12th December 2019 General Election results!

Using an exclusive algorithm, TMS is able to extract information from numerous sources, including electronic documents, scientific publications and social media. This blog post is the start of a series, which will use the General Election and the presence of politics across Twitter as an exemplar case study, in order to showcase the depth and breadth of the Text Mining Solutions algorithm.

Analysis begins with the selection of key words, mentions and hashtags. Since the announcement of the December General Election, just over a week ago, Text Mining Solutions has processed over 13million tweets which included the names of political parties (in both long and short form), political mentions or hashtags. In this way, the data that TMS collects and analyses remains neutral and unbiased, with analysis taking place across the entire political and electoral space, as opposed to directing searches towards particular people.

Text Mining Solutions then processes the data within this refined selection (albeit a selection of 13million!), and puts it into an ergonomic database. This subsequently allows us to view the results in a graph format, and therefore view which political parties are being most discussed.

The level of detail TMS is able to extract allows one to analyse the data more greatly, breaking down segments further, in order to understand sentiment and how strong the feeling is.

Below is an example of simply the surface of Text Mining Solutions’ data analysis potential. This particular graph presents the most up-to-date results regarding the level discourse across Twitter of the different political parties.