I find it interesting, specially on what tools/languages they use, and how their architecture is designed to process such a big number of tweets.

Ugh, sorry, I finished reading right at "Information Sources", thinking that was the end of the article, and missed everything after that. I'll delete my previous comment, thanks.