Andy Cotgreave's Gartner BI Twitternalysis

By Ross Perez 06 Fev, 2012

Product Consultant Andy Cotgreave has been busy at the Gartner BI Summit in London today. As he watched the stream of Tweets fly by him this morning a wave of inspiration struck; it was time to viz!

More from Andy himself:

The stand has been busy all day as people seek tools to help decision making, self-service, and provide fast analytics. These seem to be the trends Gartner is pushing as the big themes of BI in 2012, and we’ve been happy to show how we can tick those boxes. We’ve spoken to a good mix of business and IT users and hope to have impressed on them all how everyone can work together, particularly with the new Data Server we released in Version 7.

The viz shows all tweets with the #gartnerbi hashtag. I used a Python script to pull the data from Twitter into a CSV and built the app from there. I like the fact that you can see the ebbs and flows of the day, especially how things are quietest when people are eating and drinking!


Submitted by Mark J. on

Andy - Could you provide the Python script? I have a programming background, but not with Python and I can see how it would be useful.


Submitted by Andy C. on

Hi Mark
I started off with this blog post from Mark Bulling:
I too have a programming background but no python skills before this exercise. I had to tweak the python script a little but the instructions were clear enough.

I have had a lot of interest in how this process was done, so I'll write up a demo/tutorial when I get time and put that up in the Community.


Submitted by Mark J. on

Nice. I did a little research and created a similar script (though not as robust) in Ruby. I added a sub-query to pull more user info (like name and # of followers) through a user query, but I'm currently hitting the rate limit of 150 queries with Twitter for some reason, even with a small result size. With a little troubleshooting, I should be able to get it to work. What would be really interesting would be to add on a bayesian sentiment analysis. There are several examples using Ruby, Python, and PHP on the web.