data-science

Exploratory data analysis with pandas involves profiling datasets, inspecting missingness, examining distributions, and visualizing correlations to reveal structure before modeling.

Hypothesis testing requires choosing appropriate tests, checking assumptions, and interpreting results using effect sizes and confidence intervals to ensure rigorous statistical inference.

Prevent feature leakage during predictive modeling by applying correct train/test splitting and ensuring feature engineering and preprocessing occur only on training data.

Analyze A/B testing results by applying correct experimentation practices, checking statistical assumptions, and interpreting effect sizes and confidence intervals to derive reproducible insights.

Yes, matplotlib and seaborn are used for data visualization during EDA to profile datasets, examine distributions, and visualize correlations before predictive modeling.

Predictive models perform poorly on new data when feature leakage occurs during preprocessing, or when train/test splitting is incorrect, violating modeling guardrails.