Why Tweet Analyzer - doraithodla/notes GitHub Wiki

I use Twitter. I am not able to keep up. I follow most of the people who follow me. I would like to pick some gems from their tweets and retweet them. But I want thesse tweets to be relevant to my audience. The broad topics are Tech and Startups. In addition I want to participate in their conversations.

The first step seems to me segmenting users Who writes about what? TF/IDF and k-means spring to my mind. This should roughly segment tweets (the last 100). Another way could be to look at their profiles and segment. I am not sure about the consistency between the profile description and their tweet conversations.

I see the following tasks.

1. Get a list of my followers and the people I follow (including from lists)
2. Gather profiles
3. Do a TF/IDF and Topic Modeling using LDA
4. Fine tune the topics through iteration 5.LDA may not be the most current so find the current research in this space

As much as possible do it with open source code and existing Python modules.

1. Repeat the topics with the most recent 20/100 tweets.
2. Check for alignment
3. Find the Twitter influencers
4. Find the most popular topics
5. Track their trends

Follow the people and conversations which are most relevant to my needs. So perhaps a different twitter client!