JosephineKa/Malawi-Travel-Cost-Predictor-

Machine learning model approach to predict trip costs to Malawi using Polynomial Regression # 🇲🇼 Malawi Travel Cost Predictor > A machine learning system that predicts total trip costs for travelers to Malawi with near-perfect accuracy (R² = 1.000) ## 📋 Overview Planning a trip to Malawi involves navigating complex and variable costs, including flights, accommodation, and tour packages. This project develops a **machine learning-based predictive tool** that analyzes historical travel data to accurately estimate total trip expenses. Using a synthetically generated dataset of 100 tour records derived from official Malawi tourism pricing sources, we implemented and evaluated three regression models: - **Linear Regression** (Baseline) - **Lasso Regression** (Feature selection) - **Polynomial Regression Degree 2** (Best performer) ### Key Results | Model | R² Score | RMSE | MAE | |-------|----------|------|-----| | Linear Regression | 0.48 | $860.13 | $2380.59 | | Lasso Regression | 0.94 | $262.30 | $12.73 | | **Polynomial Regression** | **1.00** | **~$0** | **~$0** | --- ## 🎯 Problem Statement Travelers and tourism professionals face difficulty predicting total expenses due to: - Multiple interacting cost factors (flights, hotels, tours) - Seasonal price variations - Complex package deal interactions - Non-linear pricing structures **Our Solution:** An ML-powered prediction tool that captures non-linear relationships to provide accurate cost estimates. --- ## 📊 Dataset ### Data Sources Due to limited real-world data, a synthetic dataset was generated from three official tourism sources: | Source File | Records | Description | |-------------|---------|-------------| | `airlinesdates11.xlsx` | 60 | Airline routes, flight times, and costs | | `allregions.xlsx` | 25 | Accommodation rates and details | | `specificplace1.csv` | 22 | Tour package information | ### Generated Dataset - **100 synthetic tour records** created by combining random samples from each source - **10 categorical features** ( …