Skip to content

DanielWidrich/FoodHub-Demand-and-Customer-Behavior-Analysis

Repository files navigation

FoodHub-Demand-and-Customer-Behavior-Analysis

Exploratory Data Analysis for Marketplace Optimization

Business Problem

FoodHub is an online food delivery aggregator operating in New York City, connecting customers with a large number of local restaurants through a single mobile application. As competition among restaurants and delivery platforms increases, FoodHub needs to better understand customer demand patterns, restaurant performance, and operational bottlenecks in order to improve customer experience and drive revenue growth.

The company seeks to answer key questions around when customers place orders, which restaurants and cuisines drive the majority of revenue, how customer retention behaves over time, and whether delivery performance differs across operating conditions. Insights from this analysis are intended to inform marketing strategy, restaurant partnerships, customer retention initiatives, and delivery operations planning.


Forensic Research & Key Insights

Exploratory analysis of historical order data revealed several strong and actionable patterns:

  • Severe demand imbalance between weekdays and weekends
    Weekend orders average approximately 675 orders per day, compared to 109 orders per weekday, representing over a 600% increase in weekend demand. Weekends account for roughly 71% of total order volume.

  • Delivery performance differs by day type
    Despite higher order volume, weekend delivery times are faster than weekday delivery times, suggesting weekday congestion effects such as lunch-hour traffic or operational constraints.

  • Revenue concentration among a small subset of restaurants
    Approximately 52% of total revenue is generated by just 20 restaurants, with the top 10% of restaurants accounting for over 60% of all orders. This indicates a highly concentrated marketplace dynamic.

  • Cuisine demand is highly skewed
    Four cuisines — American, Japanese, Italian, and Chinese — collectively drive over 82% of all orders, highlighting clear customer preferences.

  • Low customer retention
    A significant majority of users are infrequent customers:

    • 65.33% placed only one order
    • 22.25% placed two orders
    • Only 1.15% placed five or more orders
      This indicates a substantial opportunity for improving repeat usage.
  • Rating behavior is polarized
    No ratings below 3 were observed. Most customers either leave no rating (38.78%) or give a 5-star rating, suggesting friction in the feedback process rather than dissatisfaction.


Modeling & Analysis Approach

This project focused on exploratory data analysis and business insight generation rather than predictive modeling. Analysis was conducted using Python with pandas, NumPy, and visualization libraries to examine demand patterns, customer behavior, restaurant performance, and operational metrics such as preparation and delivery time.

The emphasis was placed on interpreting patterns that directly influence business decisions, rather than optimizing statistical metrics.


Business Recommendations

Based on the findings, several strategic recommendations emerge:

  • Develop strategic restaurant partnerships
    Strengthen relationships with top-performing restaurants through exclusive promotions, featured placement, or co-marketing initiatives.

  • Focus marketing around high-demand cuisines
    Concentrate promotional campaigns and bundle offers around the four dominant cuisines to maximize conversion rates.

  • Improve customer retention
    Introduce incentives for second and third orders, such as targeted discounts, loyalty rewards, or personalized reminders to reduce churn after first use.

  • Increase weekday engagement
    Launch weekday-only promotions, lunch-hour deals, or corporate partnerships to offset demand imbalance.

  • Address weekday delivery delays
    Investigate time-based pricing, enhanced delivery routing, or expectation management via improved ETA transparency.

  • Enhance feedback collection
    Simplify the in-app rating process and encourage feedback participation to support quality monitoring and referrals.


Repository Structure

  • Widrich_FoodHub_Analysis_Full_Code.ipynb
    Complete exploratory data analysis, visualizations, and insights.

Tools & Technologies

  • Python
  • pandas, NumPy
  • Data visualization libraries (Matplotlib / Seaborn)
  • Jupyter Notebook

About

End-to-end exploratory data analysis of food delivery demand, customer behavior, and revenue concentration to support marketplace optimization.

Topics

Resources

Stars

Watchers

Forks

Releases

No releases published

Packages

 
 
 

Contributors