🚁 Drone Battery Performance Analyzer

A machine learning-powered dashboard for analyzing and predicting drone battery performance based on flight parameters. Shout-out to https://www.kaggle.com/samsudeenashad for this dataset: https://www.kaggle.com/datasets/samsudeenashad/supplemental-drone-telemetry-data-and-operations-lo

Key Features

• Exploratory Data Analysis (EDA)

  • Interactive visualizations of drone flight data
  • Correlation analysis between flight parameters and battery performance
  • Distribution analysis of battery levels across different drone models

• Machine Learning Models

  • Multiple regression models (Random Forest, XGBoost, Linear Regression, etc.)
  • Model performance comparison (R2 scores, RMSE)
  • Cross-validated predictions

• Interactive Dashboard

  • Filter data by drone model, application, payload type, and flight status
  • Custom flight parameter inputs for battery prediction
  • Side-by-side model comparison
  • Interactive visualizations with Plotly

Getting Started

Prerequisites

• Python 3.8+
• pip package manager

Installation

1. Clone this repository
2. Install dependencies:

pip install -r requirements.txt

Usage

1. Running the EDA & Modeling Script

python eda_and_model.py

This will:

• Load and clean the drone telemetry data
• Perform exploratory data analysis
• Train and evaluate multiple machine learning models
• Save trained models to all_models.pkl

2. Launching the Dashboard

streamlit run app.py

The dashboard provides:

• Interactive data exploration tabs
• Model selection and comparison
• Battery prediction form

Using the Prediction Form

1. Select models to compare from the sidebar
2. Enter flight parameters:
   • Duration (minutes)
   • Altitude (meters)
   • Distance (km)
   • Weight ratio (Actual/Max)
   • Wind speed (m/s)
   • Obstacles encountered
3. Click "Predict Battery" to see results

Data Description

The dataset contains drone flight records with:

• Flight parameters (duration, altitude, distance)
• Payload information
• Environmental conditions
• Battery performance metrics
• Over 500 flight records across multiple drone models

License

This project is licensed under the MIT License - see the LICENSE file for details.