This is the code for a simple Spyre web application which conducts text analysis on tweets of the 2016 US presidential candidates.
You can see this web app in action on my website, and read about how it was made on my blog.
Find out more about Spyre here.
This was my fourth project for the Metis data science bootcamp.
####Technologies used:
- Python
- NLP (natural language processing) for sentiment analysis
- Spyre
- Amazon Web Services
- tmux
In addition, I explored other technologies which didn't make it into the final implementation of the app including:
- MongoDB
- Topic modeling with LDA (Latent Dirichlet Allocation)
- word2vec
- Clustering (kmeans, DBSCAN, and others)
- Heroku