Solar Radiation Prediction

Project Overview

This project aims to predict solar radiation levels based on various environmental and atmospheric factors. By analyzing features such as temperature, pressure, humidity, wind direction and speed, and time-based metrics, the goal is to develop a regression model that can accurately forecast solar radiation. This is a crucial task for solar energy systems, weather forecasting, and agricultural planning.

---

Technical Highlights

• Dataset: Kaggle - SolarEnergy
• Size: 32,686 entries, 11 columns
• Key Features:
  • Temperature, Pressure, Humidity, WindDirection(Degrees), Speed.
  • Time-based Features: Time_Hour, Time_Minute, Time_Second, SunriseHour, SunriseMinute, SunsetHour, SunsetMinute.
• Approach:
  • Data Cleaning: The Data, Time, TimeSunRise, and TimeSunSet columns were used to engineer new time-based features and then dropped.
  • Exploratory Data Analysis: A variety of visualizations, including line plots, histograms, box plots, and a pairplot, were used to analyze data distributions and feature relationships. A bar chart of average radiation by hour highlighted a clear temporal pattern.
  • Regression Task: The target variable is Radiation.
  • Models Used:
    • A suite of regression models were trained, including Ridge, XGBoost, Random Forest, AdaBoost, Gradient Boosting, Bagging, Decision Tree, SVR, and K-Nearest Neighbors (KNN).
• Best R² Score:
  • 0.944 with KNN Regressor.
  • 0.943 with Random Forest Regressor.
  • 0.933 with Bagging Regressor.
  • The high R² scores indicate that the models are highly effective at predicting solar radiation based on the combination of atmospheric conditions and time-based features.

---

Purpose and Applications

• Solar Energy Forecasting: Enables accurate prediction of solar power generation for energy grid management.
• Agricultural Planning: Assists farmers in optimizing planting and irrigation schedules based on expected solar radiation.
• Weather Forecasting: Provides a foundational model for more complex meteorological analysis.
• Environmental Monitoring: Supports research on the effects of solar radiation on environmental systems.

---

Installation

Clone the repository and extract the data from the zip file.

Install the necessary libraries:

pip install pandas numpy seaborn matplotlib scikit-learn xgboost

---

Collaboration

We welcome contributions to improve the project. You can help by:

• Performing comprehensive hyperparameter tuning for the top-performing regression models.
• Investigating the impact of the data cleaning and feature engineering steps, and trying alternative methods.
• Exploring more advanced time-series forecasting techniques that can capture temporal dependencies.
• Adding explainability (e.g., SHAP or LIME) to understand which factors are the most significant drivers of solar radiation.