2026
Statistical inference with chi-square, ANOVA and mutual information
Hypothesis tests applied to telecom customer churn and to the relationship between continuous and categorical variables, checking assumptions before each test.
Problem
Determine which variables are truly related to an outcome instead of assuming it.
Data
Telco Customer Churn (7,043 customers, categorical features and `churn`), a color-preference survey with three alternative outcomes, and a course-provided `anova_05.csv` (5,000 rows, 4 continuous variables and an A/B/C category). The files are not included.
Method
Chi-square tests of independence at α = 0.05; mutual information; ANOVA only where normality (Shapiro-Wilk) and equal variances (Levene) hold.
Result
Having dependents (χ² = 189.1) and a partner (χ² = 158.7) are associated with churn, while gender is not (p = 0.49). ANOVA was applied only to the one variable that met its assumptions.
Stack
- Python
- SciPy
- pandas
- scikit-learn
- Seaborn
Context
Course: Data Mining · 2026