Slyth3/Sentiment-analysis-of-South-African-Banks-POC

Sentiment Analysis on the top 4 banks in South Africa # Bank Sentiment Analysis POC This project is to prove the concept of scraping the required tweets, use an out-the-box model to determine sentiment of tweets and visualize/analyse the results Process: * Twint to scrape tweets of the top 4 banks in South Africa Twint (github.com) * Clean tweets with WordPunctTokenizer and Regex * TextBlog to process sentiment of tweets * Matplotlib / Seaborn to visualise data Any tweets referencing the top 4 South African banks are scraped and their sentiment scored as eeher postive, neutral or negative: - Standard bank - Absa - Nedbank - FNB Outfiles can be found on AWS S3 (as pickle files: * pre-cleaned file (twint-nlp-banks.s3.eu-west-…) * pre-cleaned CSV file (twint-nlp-banks.s3.eu-west-…) * Final output, cleaned and with sentiment analysis (twint-nlp-banks.s3.eu-west-…)