Social Media Data Mining

TWITTER FIRE HOSE

Sifferkoll has developed a platform for analyzing selected parts of the Twitter fire hose feed. This includes the complete feed using the Swedish language, all tweets containing S&P 500 cashtags and everything with a relation to the energy market; ie. LENR, global warming, climate change, oil & gas, solar and other alternative energy technologies.

The platform can be used to analyze trends and sentiment in those areas. If this is interesting to you, use the form below to get in touch.

TWITTER CASHTAG ANALYZIS

Below are the most frequently used cashtags on Twitter during the latest week (size) and the latest day change (color)

This is what my political sentiment AI algo looked like analyzing Twitter data from the week prior to the election. It actually amazes me how close to the actual results it is. Check the two “sentiment” bars atopinion1 the bottom of the graph.

The most important points to take home is that the Sifferkoll AI-algo actually made better predictions of the outcome than most polls. Especially in the politically incorrect realm. Traditional media and polls vastly overestimated FI and MP and underestimated SD. Probably partly because of journalistic bias and their view of good versus bad. We were supposed to believe that MP and FI were active in the social medias and that SD were not, when it was clearly the other way around.

PR agency Springtime report.

Example of a extremely biased reception in media (DN).

This is an example of how the political opinion has developed between february and september 2015.

opinion2

And here is an example report for the week ending september 11, 2015

opinion3

If you instead have an interest in unbiased real time social media analytics you can contact me by filling out the form below.

Här finns lite mer utförlig statistik över det politiska läget på svenska Twitter.

 

Swedish Election on Twitter in Real Time

An experiment in data mining using a twitter live feed to get real time data from the twitter API. Political sentiment is analysed using a proprietary AI algorithm.

 

 

1 thought on “Social Media Data Mining

  1. Dear Sifferkoll Team,

    I’m a Master Student of Industrial Engineering in field of System Management and Productivity at SRBIAU (Islamic Azad University, Science and Research Branch; Tehran-Iran). I am working on my thesis and it is about real time data analysis of social networks such as Twitter, Instagram and Facebook. Therefore, I have a question: Does your product support data in Persian language?

Leave a Reply

Your email address will not be published. Required fields are marked *