Apple Tweets Sentiment Analysis (R)

Analyze public sentiment around Apple using a reproducible R workflow (text preprocessing, word frequencies, word clouds, and NRC lexicon sentiment).

Tech stack: R, tm, syuzhet, ggplot2, wordcloud/wordcloud2, dplyr, SnowballC

🎯 Project goals

• Clean and normalize tweet text
• Visualize frequent terms (bar chart, word cloud)
• Score emotions & polarity with the NRC lexicon
• Export publication-ready figures to figures/

📁 Repository structure

apple-sentiment-analytics/
├─ data/                # Put apple.csv here (see below)
├─ figures/             # Output charts saved here
├─ src/
│  └─ sentiment_analysis.R
├─ install.R            # One-time package installer
├─ LICENSE              # MIT
└─ README.md

📦 Setup

1. Install dependencies (first run only):

source("install.R")

2. Add the dataset:

• Download apple.csv from a public source like Kaggle (e.g. Apple Tweets dataset used in tutorials).
• Place the file at: data/apple.csv

If you use someone else’s dataset or tutorial for inspiration, please credit the source in this README.

3. Run the analysis:

source("src/sentiment_analysis.R")

This will generate:

• figures/frequent_words_barplot.png
• figures/wordcloud.png
• figures/sentiment_scores_barplot.png

🖼️ Figures

View all figures ➜

Frequent Words (Barplot)

Word Cloud

NRC Sentiment Totals

Top Positive vs Negative Words (bing lexicon)

📝 Method overview

• Text cleaning: lowercase, remove punctuation/numbers/URLs/stopwords; stemming; custom domain terms removed (aapl, apple).
• Term frequency: Term-Document Matrix via tm; top terms visualized.
• Sentiment: syuzhet::get_nrc_sentiment() for 10 emotion/polarity categories.
• Visuals: Base R + ggplot2 + wordcloud/wordcloud2.

🔍 Key insights (example from my run)

• Discourse clusters around earnings, reports, and stock movement.
• Mixed sentiment: high negative but also strong anticipation and trust, reflecting market tension.

🤝 Attribution

Inspired by community tutorials on text mining in R. If you adapt from a specific notebook (e.g., Kaggle), please add a link here and describe what you changed.

⚖️ License

MIT — see LICENSE.

---

Author: Eva Samitova
Course: Text Analytics
Date: 2025-07