Big-Box Site Study Big-Box Site Study

Overview

Delivered a multi-format site study for a Latin American home-improvement retailer's prospective big-box store in the Monterrey metro area, combining ten documented models, from household spending baskets and pass-by mobility to origin–destination capture and cannibalization, in one auditable web report, deck and data package.

Highlights

1. 01

   Built read-only Python models over census, NSE, ENIGH 2024 spending, DENUE on H3, device mobility and a 21M-row origin–destination matrix. Each model emits a summary the integrator only copies, so no figure is typed by hand.

2. 02

   Measured cannibalization against the chain's existing stores with a layered ladder (geometric overlap, O-D origin overlap, co-visits, drive-time convenience) and calibrated pass-by traffic with site-specific NSE scaling.

3. 03

   Shipped an interactive report with 16 GeoJSON layers, a 33-slide deck, a PDF, an XLSX and a data ZIP, gated by build-failing validators after a 47-fix expert review.

Stack

• Python
• PostgreSQL
• PostGIS
• H3
• Origin–Destination Analysis
• Huff Gravity Model
• Isochrones
• ENIGH/AMAI NSE
• Vite
• Leaflet
• Node.js Validators
• PPTX/PDF Automation