In the vibrant realm of social media, where emoji’s tell tales and hashtags wield power, this article embarks on an expedition into the heart of sentiment surrounding the South African Rand. With an arsenal of tweets and the guidance of Natural Language Processing (NLP), I journey through the digital tapestry, seeking to decipher the intricate choreography between sentiment and currency.
Summary
- Sampling Methodology: My quest commences by delving into the meticulous art of sampling, capturing the diverse chorus of tweets that resonate with the Rand’s essence.
- Frequency Analysis: The narrative then turns to frequency analysis of these tweets, unearthing patterns that paint a portrait of public perception.
- Word Cloud: A mesmerizing word cloud emerges, weaving a vivid narrative of the lexicon surrounding the Rand.
- Quantifying Sentiment: Navigating the realm of sentiment quantification, FinBERT steps forth as our guide, assigning sentiment values to each tweet.
- Sentiment over Time: The Twitter Rand Sentiment Index (TRSI) emerges, mirroring the rhythm of sentiment across time, inviting exploration into its nuances.
- TRSI and the Exchange Rate: Exploring the potential of the TRSI as a guide for exchange rate insights
- Future Analysis and Refinements: The curtain falls on this chapter, yet the stage is set for future analytical endeavors. Expanding horizons, masterful pre-processing, and innovative feature engineering stand as pillars of exploration, poised to enrich the narrative.
The rest of this article is available on Medium and Linkedin