Team Fast Fashion

Inspiration

Noticed for luxury apparel how customer engagement is heavily focused on Instagram

What it does

Take sentiment analysis and engagement metrics from Quandl's social media dataset to produce a buy/hold/sell recommendation based on a rolling 90 day window

How we built it

Harvested data from quandl with EOD prices pulled manually from Quantopian. Model training done via Keras. Backtesting strategy via Quantopian

Challenges we ran into

Short time window of facebook data. Limited to S&P 500 companies where the strategy is more applicable to smaller cap companies

What we learned

None of us had exposure to analysis with alt data

Built With

• keras
• quandl