Elections Polls are based on sampling a portion of the voting public which is incomplete NaMoStats was inspired to provide a holistic view to the voting public to know the candidates better

What it does

Showcase spatial & temporal trends of twitter activity of candidates, their mentions, people's view

How we built it

We used twitter API to get data. We used open source tools like solr to create and analyse the content Used google maps api and D3 to visualize the analytics. We also created a promptapp command based interface.

Challenges we ran into

Indexing & Enriching our Solr Index was time-consuming despite performing Atomic Updates. We were limited by our AWS nano account

Accomplishments that we're proud of

We were able to visualize temporal & spatial trends in Tweets

What we learned

What's next for NaMoStats-Analyzing 2016 US Presidential Elections

Grab and parse data from Facebook, news sites to enrich our Solr index Run LDA Topic Modeling on URLs being shared by candidates Compute clustering of left wing & right wing followers based on affinity to Candidates

Share this project: