Analysis methodology
A reproducible workflow transforms the Restaurant Tips dataset into findings that are clear enough to support operational decisions.
Before analysis
Meaningful questions
- 1.What is the typical total bill, tip amount, and tip percentage?
- 2.Do lunch and dinner guests have different spending and tipping patterns?
- 3.Which service days show the highest activity and average bills?
- 4.How does party size relate to bill value and tip amount?
- 5.Are there unusually high bills or unusually low tip percentages?
Cleaning process
Quality controls
- Standardize variable names and categorical text.
- Check missing values, duplicates, and numeric validity.
- Remove invalid non-positive bills and negative tips.
- Create tip percentage as a derived analytical feature.
- Use the IQR rule to flag unusual bill values.
Statistical analysis and reporting
Evidence beyond visual impressions
The Python pipeline creates descriptive statistics, grouped summaries, correlation analysis, anomaly checks, and a Welch independent-samples t-test.
Descriptive analysis
Mean, median, spread, minimum, maximum, and grouped metrics establish the overall data profile.
Hypothesis test
A Welch t-test evaluates whether mean dinner and lunch tips differ significantly at alpha = 0.05.
Visual story
Histograms, boxplots, scatter plots, bar charts, and a heatmap make patterns and outliers understandable.