WorldPulse Data Explorer

An interactive Streamlit app to explore, profile, manipulate and visualize WorldPulse datasets. Built in collaboration with the WeSkill volunteering program.

Features

• Home: Overview and instructions.
• Data Upload: Upload CSV/Excel files and preview with an interactive grid.
• Manipulation: Launch a D-Tale session for ad-hoc data exploration.
• Profiling: Generate an automated Pandas Profiling report.
• Visualization: Create custom charts with PyGWalker and AutoViz.
• AutoML: Train and compare classification/regression models via PyCaret.
• Preprocessing: One-hot, target encoding, normalization, etc.

Getting Started

# 1. Clone the repo
git clone https://github.com/equinix-labs/worldpulse-data-explorer.git
cd worldpulse-data-explorer

# 2. (Optional) Create & activate a virtual environment
python3 -m venv .venv
source .venv/bin/activate

# 3. Install dependencies
# If using Python 3.13+, use ignore-requires-python to bypass unmet python_requires constraints
pip install --ignore-requires-python -r requirements.txt

# 4. Run the app
streamlit run WorldPulseMain.py

The app will be accessible at http://localhost:8501 in your web browser.

Project Structure

worldpulse-data-explorer/
├── WorldPulseMain.py    # App entrypoint
├── home.py             # Home page
├── data_upload.py      # File upload & AgGrid preview
├── manipulation.py     # D-Tale integration
├── profiling.py        # Pandas Profiling
├── visualization.py    # PyGWalker charts
├── autoviz.py          # AutoViz charts
├── encoding.py         # Encoding utilities
├── preprocessing.py    # Scaling / normalization
├── regression_pycaret.py # Model training UI
├── requirements.txt
└── README.md