nazjobs/solar-data-discovery

Analysis of West African solar potential with an interactive Streamlit dashboard. # Solar Data Discovery and Analysis for West Africa **A Week 0 Project for the KIFIYA - Kaim 10 Academy Generative AI Engineering Program.** ## Project Overview The objective of this project is to analyze solar irradiance and weather data for three potential locations in West Africa—Malanville (Benin), Bumbuna (Sierra Leone), and Dapaong (Togo)—to provide a strategic recommendation for the optimal placement of a new solar energy farm. The analysis follows a structured data science workflow, from initial setup and data cleaning to in-depth exploratory data analysis and statistical comparison. --- ## 🚀 Live Demo An interactive dashboard for this project has been deployed to the Streamlit Community Cloud. It allows for a visual and quantitative comparison of the solar potential across the three locations. **Click here to view the live application.** ---- ## Analysis & Key Features * **Professional Environment Setup:** The project is configured with a reproducible Python virtual environment, with all dependencies locked in `requirements.txt`. * **Version Control:** A feature-branch Git workflow was used, with all major tasks (setup, EDA per country, comparison) isolated on their own branches and merged via Pull Requests. * **Automated CI/CD:** A GitHub Actions workflow is in place to automatically validate the environment by installing dependencies on every push and pull request. * **Data Cleaning & Profiling:** Each country's dataset (~525,000+ rows) was systematically profiled. Key cleaning steps included: * Handling of physically impossible negative solar irradiance values. * Removal of empty or useless columns. * Statistical flagging of outliers using Z-scores. * **Exploratory Data Analysis (EDA):** Individual notebooks were created for each country to explore: * Time series patterns of solar irradiance (GHI). * Distributions of key weather variables like ambient temperature and wind speed. * Correlation analysis between variables usi …