Solar-ML-Project

Machine Learning model for predicting solar power generation using weather and time-series features.

Project Overview

This project aims to build a machine learning pipeline that predicts solar energy output based on environmental variables such as: 1.Solar irradiance 2.Temperature 3.Humidity 4.Time of day 5.Weather conditions The goal is to help optimize solar panel efficiency and forecast power generation.

Tech Stack

1.Python 2.NumPy 3.Pandas 4.Matplotlib / Seaborn 5.Scikit-Learn 6.Jupiter Notebook 7.Git / GitHub

Project Structure

Solar-ML-Project

│── data → Raw or cleaned datasets
│── notebooks → Jupyter/XLS models
│── models → Saved ML models
│── src → Python training scripts
│── README.md → Project documentation
│── requirements.txt → Dependencies

Workflow

1️⃣ Data Collection Collect solar weather data (CSV or API): 1.Solar radiation 2.Temperature 3.Humidity 4.Wind speed 5.Time features 2️⃣ Data Preprocessing i.Handle missing values ii.Normalize numerical features iii.Encode categorical features 3️⃣ Model Training

Models used:

1.Linear Regression 2.Random Forest Regressor 3.XGBoost Regressor (optional) 4️⃣ Model Evaluation

Metrics:

MAE

RMSE

R² Score

5️⃣ Deployment (Optional)

Simple Flask API or Streamlit dashboard.

How to Run This Project

Install dependencies:

pip install -r requirements.txt

Run Jupyter Notebook:

jupyter notebook

Train the model:

python src/train.py

Results

Random Forest achieved the best accuracy

Model stable under varying sunlight intensity

Produces reliable short-term predictions

(Add graphs/screenshots later)

Future Improvements

Add Deep Learning (LSTM) for time-series prediction

Real-time data streaming API

Build a web dashboard

Contributions

Pull requests are welcome! For major changes, please open an issue first.

Author

Saiful Islam Mahin