Recipes & Ratings Analysis 🍽️

Exploring the Sweet Science of Recipe Happiness
🍦🍨🍰🍜🍕🍔🌮🥘
Authors: Luran Zhang, Yunpeng Zhao
View on GitHub

Welcome to Our Data Analysis Journey!

We've analyzed over 83,782 recipes from Food.com to uncover fascinating insights about what makes people happy in the kitchen. Our research combines statistical analysis with machine learning to answer two compelling questions about recipe success and nutritional prediction.

📊 Project Overview

83,782 Recipes Analyzed
2 Main Analyses
84.9% Model Accuracy
2008+ Data Range

🔬 Our Research Questions

🧐
More Sugar == More Happiness?
Investigate the relationship between sugar content and recipe ratings to understand if sweet treats truly bring joy to people's lives.
  • Statistical hypothesis testing
  • Missingness analysis
  • Interactive visualizations
  • Comprehensive data cleaning
Explore Analysis →
🤤
Guess the Calories
Build a machine learning model to predict recipe calories using nutritional features and recipe characteristics.
  • Decision Tree Regression
  • Feature engineering
  • Fairness analysis
  • Model validation
Explore Analysis →

🎯 Key Insights

  • Sugar vs. Happiness: No significant correlation found between sugar content and recipe ratings (p-value = 0.13)
  • Calorie Prediction: Our machine learning model achieves 84.9% accuracy on test data
  • Model Fairness: No evidence of bias between old and new recipes (p-value = 0.26)
  • Data Quality: Missingness patterns reveal interesting temporal trends in recipe submissions