1 minute read

This project analyzes delivery order data from Cincinnati over a 30-day period, spanning September 16 to October 14, 2022. The dataset contains 60,583 entries across 19 columns, covering order details, delivery timestamps, driver assignments, and grocery partner performance.

🚨 Executive Summary

Our service is facing two critical challenges:

  • Late Deliveries (4.7%) — driven by extreme delays in driver acceptance, especially for low-priced orders
  • Order Accuracy (2.5%) — compromised by grocery store partners, leading to unresolved complaints
  1. Restructure Driver Incentives — make low-priced orders more attractive
  2. Partner Success Program — deploy DashMart’s best practices across grocery partners
  3. Peak Hour Optimization — boost driver supply from 19:00 to 00:00

👉To reproduce this on the dataset or a similar dataset, you can follow the Jupyter Notebook


📊 Data Quality

  • Data Quality: High, with minor completeness issues
  • Low percentage of missing data: DELIV_DASHER_ID, DELIV_D2R, DELIV_CLAT . These gaps are unlikely to affect most analyses. Therefore, I’ve chosen to drop null rows, as these fields are essential for downstream processing
  • Intentional Blanks: DELIV_CANCELLED_AT, SUBSTITUTE_ITEM_NAME, CATEGORY

🔥 Market snapshot

  • Total Orders: 60,583 (Sep 16–Oct 14, 2022, Cincinnati)
  • Cancellation Rate: 1.1%
  • Late Delivery Rate: 4.7%
  • Missing/Incorrect Rate: 2.5%

  • Order Peaks: 23:00–03:00 and 18:00–22:59
  • Stable Hours: 06:00–11:00
  • Critical Window: 19:00–00:00 — highest concentration of late and missing deliveries
  • Cancellations: Low and consistent

Hourly Heatmap


🧍 Customer Experience

Order Accuracy:

  • Resolution Gap: Only 8% of complaints were resolved with substitutions
  • Mismatch Rate: 166 substitute items didn’t match original category

Substitution Accuracy

Wait Time:

  • Median Acceptance Time: ~1.5 minutes
  • Outliers: Up to 120 minutes
  • Insights: Low-priced orders (< $0.20) ignored by drivers

Acceptance Time Distribution


🚗 Dasher Performance

  • Travel Time Impact: Longer store arrival → higher chance of late delivery
    Travel Time vs Late Status

  • Price Sensitivity:

    • Orders < $0.20 → longest wait times
    • Orders > $0.40 → fast, predictable acceptance

Price vs Acceptance Time Distance vs Acceptance Time


🛒 Merchant Reliability

DashMart vs Grocery Stores:

  • DashMart:
    • Missing Rate: 0.2%
    • Substitution Rate: 0.1%
  • Grocery1:
    • Missing Rate: 14.8% Merchant performance
  • Share top 5 items per top 5 categories
  • Help merchants stock popular items and improve substitutions

Top Items Bubble Chart


🧠 Strategic Recommendations

1. Fix the Small Order Problem

  • Batching: Combine small orders
  • Minimum Fare Guarantee
  • Wait-Time Bonuses

2. Transform Grocery Partnerships

  • Short-Term: Launch Partner Success Program
  • Long-Term: Shift toward DashMart’s scalable model

3. Reinforce Peak Hour Operations

  • Driver Supply Surge: Incentivize 19:00–00:00
  • Push Notifications: “Evening Rush Bonus”