Urban commutes are more than just a daily hassle—they are a data problem. This project analyzes survey responses to move past "gut feelings" and prove exactly what is clogging our roads.
- Data Wrangling: Pandas for mode imputation and ordinal encoding.
- Visualization: Seaborn & Matplotlib for distribution analysis and proportional stacked bars.
- Statistics: Scipy for Chi-Square Contingency tests to validate urban hypotheses.
-
Infrastructure is the Enemy: A definitive, non-random link exists between "Poor Road Infrastructure" and high-duration traffic delays (
$p < 0.001$ ). -
The Generational Shift: Younger commuters (18-45) are statistically more likely to switch to public transit if reliability improves (
$p = 0.041$ ). -
Mode Matters: Primary commute modes (Public vs. Private) show significantly different congestion frequency patterns (
$p = 0.021$ ).
- Cleaning: Stripping whitespace, renaming columns, and handling missing values with mode imputation.
- Transformation: Ordinal encoding of categorical features like
AgeandCommuteTime. - EDA: Visualizing distributions across residential areas and commute modes.
- Testing: Statistical validation of causes vs. extra time added by congestion.
- Targeted Infrastructure: Improving road quality directly alleviates long-duration congestion.
- Policy Focus: Launch transit reliability campaigns specifically targeting the 18–45 demographic to maximize adoption.
Made with ❤️ by Manas Shukla
