Solar Energy Forecasting Dashboard

This project explores how we can use open energy data and forecasting models to better understand short-term trends in solar power generation. The goal was to build a simple but useful dashboard that shows both historical and predicted solar output, using real hourly weather data from the UK.

📌 Project Motivation

As interest in renewable energy grows, the ability to forecast solar generation becomes more important — for grid operators, policymakers, and sustainability teams alike. I wanted to explore how a data analyst could contribute to this problem using publicly available data, Python, and Power BI.

This project was also a way for me to practice working end-to-end: cleaning real-world data, building a forecasting model, and presenting the results in a clear, visual format.

🛠 Tools & Methods

• Data Source: Open EU/UK solar energy dataset (hourly)
• Modeling: Facebook Prophet for time-series forecasting
• Cleaning:
  • Filtered out zero or negative solar generation values
  • Parsed datetime and clipped top 5% of values to reduce outliers
• Forecasting:
  • Trained model on historical hourly data
  • Forecasted solar output for the next 7 days
• Visualization: Built a two-page Power BI report to compare actual and forecasted trends

📊 Dashboard Overview

The Power BI dashboard includes:

• A time-series line chart comparing historical and forecasted energy output
• Clear axis labels, legend, and layout across two report pages
• A simple design focused on readability and storytelling

💡 Key Insights

• The model captured the general pattern of daily solar output, including typical rises and falls across hours.
• Forecasting works best when data is clean and focused — trimming extreme outliers helped reduce noise.
• Power BI allowed for a straightforward way to present both historical and forecasted trends side by side.

📁 Files Included

• solar_data_raw.csv – Original dataset (UK hourly energy + weather)
• solar_data_cleaned.csv – Preprocessed version for Prophet
• solar_forecast_output.csv – Combined actual and forecasted results
• solar_forecasting_prophet.py – Python code for training + prediction
• solar_forecast_dashboard.pbix – Final Power BI dashboard file

🚀 What I Learned

• How to use Prophet for time-series forecasting with minimal setup
• How to prepare and clean messy real-world time data
• How to communicate results visually using Power BI

📦 Optional Dependencies

📁 To view the Power BI dashboard, download the .pbix file above and open it in Power BI Desktop.

If you want to run the forecasting script:

pip install pandas prophet matplotlib