BlackPrint Technologies · Jun 2026
Site Score Demo
Block-Level Site Selection & Sales Estimator
Overview
Built an interactive site-intelligence demo that scores 16,472 city blocks in Mérida and estimates monthly sales for a candidate branch with a log-linear regression model.
Highlights
- 01
Rendered 16,472 Mérida city blocks by a 0–100 Site Score in Leaflet, with 10 toggleable H3 layers covering 7 public competitor chains and own-format recommendation surfaces.
- 02
Implemented a log-linear OLS sales estimator (site score, log floor area, format dummies) with reverse search that highlights blocks able to hit a sales target, and reported honest validation: in-sample R² 0.92 vs leave-one-out R² 0.51 on 11 branches.
- 03
Added server-side geocoding (Photon with Nominatim fallback), saved locations, a login gate and Zod-validated data loading on Next.js 16 and React 19, with Python export scripts that regenerate the JSON layers.
Stack
- Next.js 16
- React 19
- TypeScript
- Leaflet
- h3-js
- OLS Regression
- Site Selection
- Geocoding (Photon / Nominatim)
- Zod
- Tailwind CSS 4
- Python
- Vercel
Working on something similar?
Tell me about the problem and the data behind it. I reply within 24–48 hours.
Discuss a project