Using sentiment analysis from Twitter and Reddit to figure out which cryoto currencies to buy.
-
Twitter
- Access the Twitter API.
- Get tweets by filtering with crypto-related hashtags.
- Handle API rate limits and implement pagination.
- Preprocess tweets to remove noise (URLs, mentions, etc.).
-
Reddit
- Access the Reddit API.
- Get posts from crypto-focused subreddits.
- Handle API rate limits and implement pagination.
- Preprocess posts to improve sentiment analysis accuracy.
-
Sentiment Analysis
- Use the TextBlob sentiment analysis tool.
- Analyze sentiment of tweets and posts.
- Aggregate sentiment scores for each cryptocurrency.
-
Data Visualization
- Create interactive dashboards with Plotly/Dash.
- Display sentiment trends over time.
- Visualize sentiment distribution among different cryptocurrencies.
-
Access Binance API
- Fetch current prices of top 10 cryptocurrencies identified by sentiment analysis.
- Compare sentiment analysis results with historical price data.
- Implement robust error handling for API interactions.
- Consider data privacy and ethical aspects when presenting user-generated content.
- Plan for scalability and transitioning from a notebook to a more user-friendly application.
- Explore incorporating additional data sources like news articles in future versions.
- Investigate custom machine learning models tailored to cryptocurrency discussions.
- I was able to get a basic trading bot going in the TradingBot folder. This uses the markers from Binance to determine if a certain crypto is a buy or sell. I was also able to visualize the markers using mpfiance.