CASE STUDY

Automated ETL Pipeline For Urban Logistics

Automated ETL Pipeline For Urban Logistics
Key Business Insights & Recommendations Based on the extensive exploratory data analysis (EDA) found in the Jupyter Notebooks, several critical business insights were uncovered: 1. Surge Pricing Algorithm Optimization (Weather Impact on Delivery) Observation: There is a clear correlation between weather conditions and operational metrics. As weather shifts from 'Sunny' to 'Drizzling' and 'Raining', order volumes spike significantly (from 64.4 to 90.6 orders/hour) while average delivery times increase. The heatmap analysis proves that operations are most severely impacted when rain is coupled with moderate-to-strong winds. Business Recommendation: Implement a multi-tiered Surge Pricing algorithm. Trigger Level 1 Surge during 'Drizzling' conditions to capture the demand spike while maintaining driver supply. Trigger Level 2 Surge (maximum tier) during 'Raining' conditions with 'Moderate/Strong Winds' to compensate for the severe drop in delivery speed and higher safety risks for drivers. 2. NO2 Emissions as a Traffic Proxy: The "Photolysis" Revelation Observation: The initial hypothesis that Nitrogen Dioxide (NO2) peaks during daytime rush hours was proven false. Due to the Photolysis effect (sunlight breaking down NO2) and Thermal Inversion, NO2 levels actually plummet during the day and skyrocket at night (peaking at ~80 µg/m³ between 22:00 - 23:00). Macro-level satellite data lacks the granularity to differentiate daytime street-level traffic. Business Recommendation: Pivot from using NO2 as a traffic ETA proxy. Instead, address a critical operational risk: Driver Welfare. Night-shift logistics drivers are exposed to hazardous NO2 levels. The company should mandate and provide N95/KN95 masks as part of standard safety gear for the night fleet to mitigate long-term health liabilities. 3. The "Washout" Effect: Post-Rain PM2.5 Dynamics Observation: Heavy precipitation (> 7.5 mm) acts as a natural air purifier. The data proves a significant drop in PM2.5 concentrations immediately during and in the hours following heavy rain events. Business Recommendation: Optimize marketing pushes or delivery promos for "Clean Air Delivery Hours" immediately following rainstorms, capitalizing on higher outdoor activity intent. 4. Health-Tech Marketing Opportunity: Air Temperature vs PM2.5 Observation: There is a distinct correlation between the drop in air temperature at night and the concentration of PM2.5 particles. Business Recommendation: Implement dynamic app notifications targeting users to order food/groceries delivery during high-pollution/low-temperature evening hours, emphasizing the health benefits of staying indoors.

Other Projects you might like

View Study

HR Analytics

Jaya Jaya Maju is a multinational corporation established in 2000, with a workforce of over 1,000 employees spread across the country. Despite its significant scale, the company still faces challenges in managing its workforce. This has resulted in a high attrition rate—the ratio of staff departures relative to the total headcount—exceeding 10%. The HR Department requires assistance in identifying the various factors driving this high turnover rate and seeks a business dashboard to monitor these factors on a continuous basis.

View Study

Bank Customer Churn

This end-to-end data analytics project focuses on diagnosing and understanding customer churn within the banking sector. With a churn rate of 20% across 10,000 customers, the bank faced a critical challenge: retaining financially stable, long-term clients who were actively moving their capital to competitors. The analysis was conducted using the CRISP-DM (Cross-Industry Standard Process for Data Mining) methodology — moving beyond surface-level visualization into deep business understanding and strategic evaluation.

View Study

OTT Platform Analytics

This project explores a massive-scale user interaction dataset (approx. 14GB) from a major Video-on-Demand (OTT) platform. The primary objective is to translate raw viewing logs into actionable business strategies across four key pillars: Marketing ROI, Technical User Experience (UX), Product Feature Effectiveness, and Content Monetization. By processing millions of rows using highly memory-efficient techniques, this analysis provides clear, data-driven recommendations to optimize ad spend, reduce churn rate, and maximize platform loyalty.