Project Overview

Objective

The project aimed to optimize logistics efficiency by analyzing the geographic distribution of Walmart store locations across Illinois, Indiana, and Michigan, and their proximity to a central warehouse. The goal was to identify optimal routing clusters, streamline delivery planning, and propose data-driven strategies for scalable warehouse-store networks.

Context

In retail logistics, especially for companies like Walmart, spatial planning is crucial to reduce costs and enhance delivery speed. With over 50 stores in varying proximities to a central warehouse, understanding how to group, route, and service them is key to operational success. This project simulated a realistic logistics challenge using cleaned and curated geospatial data.

Tools and Methodologies

• Python libraries: pandas, geopy, networkx, matplotlib
• Geospatial techniques: Geodesic distance calculations, subgraph decomposition
• Future-ready extensions: CSV-based address parsing, OpenStreetMap API, clustering models

The Approach and Process

Data Collection

The raw data file included over 50 store records with fuzzy fields, incomplete formatting, and real-world-looking addresses.

To prepare the data:

• Cleaned and standardized addresses using Python string functions
• Introduced fake but realistic addresses (as location proxies) for simulation
• Defined a central warehouse in Fort Wayne, IN

Distance Computation

Using the geopy.geodesic method, pairwise distances were computed between:

• All store pairs → forming edges of the logistics graph
• Each store and the warehouse → for hub analysis

This created a fully connected weighted network graph representing physical distances.

Network Graph Construction

Used networkx to:

• Define each store as a node
• Connect stores with weighted edges
• Visualize the layout with a spring layout for interpretability

This helped understand proximity, route clustering potential, and store density.

Subgraph Analysis

To reduce complexity and enhance regional focus:

• Divided network into 3 subgraphs:
  • Stores 1–14 (Chicago metro)
  • Stores 15–35 (Michigan region)
  • Stores 36–51 (Indiana & Southern reach)