sentiment-analysis

Download / Clone from GitHub

To get the project from GitHub:

git clone https://github.com/ali-771/sentiment-analysis.git
cd sentiment-analysis

Then install dependencies using uv and setup the environment:

uv install

(Assumes pyproject.toml includes all dependencies like transformers, pyserial, sklearn, etc.)

---

# Sentiment Analysis Project

This project implements a **Sentiment Analysis System** using Hugging Face Transformers (DistilBERT).  
The model is fine-tuned on a prepared dataset and can be used both for **direct text predictions** and for **integration with ESP32** to receive messages and classify them in real-time.

---

## 📂 Project Structure

New_folder/ ├─ src/ │ ├─ data/ │ │ ├─ raw/ # raw dataset │ │ ├─ processed/ # cleaned dataset │ │ └─ clean_null.py # preprocessing script │ └─ models/ │ ├─ train_model.py # training script │ ├─ esp32.py # ESP32 bridge (serial communication) │ ├─ model_last2/ # trained model │ └─ compact_sentiment_predict.py ├─ main.py # main entry point for CLI predictions ├─ pyproject.toml # project dependencies (managed by uv) └─ uv.lock

---

## 📥 Download / Clone from GitHub

```bash
git clone https://github.com/ali-771/sentiment-analysis.git
cd sentiment-analysis

---

🚀 How to Run

1. Data Cleaning

uv run python src/data/clean_null.py

• Loads the raw dataset (sentiment_synthetic.csv).
• Removes null/empty rows.
• Saves the cleaned dataset as sentiment_synthetic_final.csv.

2. Model Training

uv run python src/models/train_model.py

• Fine-tunes DistilBERT on the processed dataset.
• Saves the trained model into src/models/model_last2/.

3. Prediction (CLI)

uv run python main.py

Example:

📝 Text: I love this
I love this  -->  positive

4. ESP32 Bridge

uv run python src/models/esp32.py

• Opens a serial connection to ESP32 (default COM10, 115200 baud).
• Reads messages from ESP32.
• Predicts sentiment using the trained model.
• Sends the prediction back to ESP32.

---

👥 Engineers:

• Ali Al-Khaledi – 202270295
• Zaid Damag – 202170119