We want to help people make decisions through providing a tool that helps enhance their extra sensory perception.

What it does

NetdB mines tweets from influential networks (influential people and who they follow). Our data pipeline then uses nlp to convert the tweets to sentiment. Market data is also downloaded and indexed into elasticsearch. Context can then be distilled through keyword searching. Filtered context sentiment and market returns are then exported to jupyter notebook for statistical testing.

How we built it

We built our nlp data pipeline with python and jupyter notebook. Searching, aggregation, and visualization are then done by Elasticsearch and Kibana. Statistical analysis is performed with Jupyter notebook.

Challenges we ran into

Twitter API has call limits and took a long time to scrape. Elasticsearch and Kibana were slow to configure for the demo because of Java path hiccups.

Accomplishments that we're proud of

Able to index our data, visual, search, aggregate, and export for statistical testing. Our statistical test reveals high significance, for example the Cointelegraph following twitter network over the last year, filtered on Coinbase context with a contrarian day trade strategy between sentiment and price resulted in a p-value <<0.005! (The unfiltered context indicated p-value=0.)

What we learned

Its important to come to the hackathon with a game plan, executing to the plan and completing goals on time is a big reward in itself, a positive outcome is icing on the cake! Know your audience for the pitch, and observe who and how people do well!

What's next for

We are going to be making moneyyyyyyyyyy!

Built With

Share this project: