Statistical Hypothesis Testing on Energy Consumption Data

This project explores the Energy Consumption Dataset using descriptive, comparative, and associative statistical methods. The analysis focuses on understanding distribution patterns, testing statistical significance of differences across groups, and evaluating relationships between variables such as average temperature, square footage, number of occupants, and number of appliances used. Key analyses include:- Distribution analysis via histograms, KDE plots, and Q-Q plots.- Normality testing using the Kolmogorov–Smirnov Test.- Wilcoxon signed-rank test to evaluate mean energy consumption assumptions.- Mann–Whitney U tests to compare energy consumption across different levels of household features.- Spearman correlation to examine non-parametric relationships. Results show that energy consumption is not normally distributed, and variables like square footage, number of occupants, and appliances used significantly influence energy usage. Meanwhile, average temperature showed no significant correlation. This project underscores the importance of checking statistical assumptions before choosing the appropriate inferential method. Tools: Pandas, Numpy, Matplotlib, Seaborn, Scipy.stats, Statsmodels, Sklearn, Kaggle CLI Dataset: govindaramsriram/energy-consumption-dataset-linear-regression Main File: train_energy_data.csv