Being from a marketing background this will come as no surprise, Social Analytics fascinate me. Having the ability to collect data, deconstruct it, analyse it and act upon findings is a key way of staying informed in a fast changing, fast developing, digital world.

We are always innovating at Triangle and invest extensively in R&D, this involves testing new products and endeavouring to come up with our own solutions. So I was particularly tuned in when it was mentioned that we were working on something in the Social Analytics space.

I have written about the things we like within IBM’s dashDB software previously, however, this time, we were looking to use the tools within dashDB for a new purpose entirely.

Flash Point

The flash point came from conversation in the office I had with my boss about trends on Social Media combined with a colleague working with a sub-feature on dashDB called “Load From Twitter” (This part of the software application provides a simple query input field where the user can type and experiment how many tweet records would be loaded with a certain query). You can define your search terms and then analyse your tweet data with Aginity which comes free as part of dashDB. So we set about loading some sample tweets.

The Technical Bit

With our gathered data we analysed a suite of tables within dashDB and ran scripts which tidied the raw data into something usable. Only then did we realise that “Load From Twitter” has an extra value add- sentiment analysis. Sentiment analysis analyses the body of the tweet and has a form of look-up or way of scoring the tweets based on a selection of words grouping these as either positive or negative, superb for those wanting to see the feedback or opinion of tweeters.

We then decided to do some reporting over the top of this data, IBM Cognos being our weapon of choice. We wrote a report which enabled the organised data to be digested in an edible format for the users.

To Put This Into Context

Using IBM Cognos, we were able to find and display trends in social media data to distinguish:

  • Popular Users and their Followers
  • Geographical information (Region,County / State, City)
  • Trends (Globally and Locally)
  • Biodata
  • Technology these tweets were entered on. (iPhone, Android, iPad, PC, Mac, etc)
  • Plus much more, it really is the case that you could interrogate the data to answer hundreds of questions.

Our Social Analytics Real World Example

Let’s say I’m a Music Executive working to bring in new talent to my record label. I want to find the next big thing before anyone else does. Aside from getting myself to gigs every night and manually trawling Social Media for hashtags and retweets, where is my data coming from to make those calls on who to invest my time with?

This simple report process could save time and money for any business who want to utilise social media to make better-informed decisions moving forward.

Posted in Analytics

January 13, 2017