parham.
All projects
2024

Retail Geospatial Analytics: Fater S.p.A. Industry Challenge

We joined Fater proprietary sales with ISTAT census in MySQL and ranked districts by per-capita store potential. I presented the work solo to Fater leadership.

Source on GitHub

20

Districts

2

Source systems

4 people

Team

Jury pick

Recognition

Mock data, real methodology

Fater's proprietary sales records are under NDA. The map below and the public repo use representative mock data that mirrors the structure of the original. The SQL queries, joins, spatial pipeline, and per-capita store-potential metric are faithful to the work I presented to leadership; only the underlying numbers are synthetic. The static deliverable that drove the jury decision used the real data and is not shown here.

Overview

An industry challenge for a Procter and Gamble joint venture

Fater S.p.A. is the P&G + Angelini Industries joint venture that owns Pampers Italia, Lines, and Tampax for the Italian market. The brief: rank 20 administrative microcode districts by expansion opportunity using a mix of their proprietary sales records and ISTAT census data that had no prior link.

Our team of four built the SQL and spatial pipeline together; I presented the final analysis solo to Fater leadership and was picked by the jury for individual recognition.

Methodology

From two disconnected sources to one ranked map

Three steps, in order:

  1. SQL stitch. CTEs and joins in MySQL combined Fater's proprietary sales records with ISTAT sociodemographic census data per district. Window functions handled the district-level rankings.
  2. Spatial join. A GeoPandas point-in-polygon pipeline (EPSG:4326) attached each store to its district polygon so per-district counts could be computed from raw store points.
  3. Per-capita metric. Store count divided by district population gives the per-capita store-potential figure that drives the ranking. Higher = under served relative to population.
Interactive map (mock data)

Choropleth of per-capita store potential

The Folium map below reproduces the shape of the deliverable that went into the Fater presentation, rendered against mock data (see the note at the top of this page). The static map export with the real data drove the jury decision; the interactive version here exists so the methodology is inspectable.

Interactive Folium choropleth of per-capita store potential

Static deliverable

Final choropleth shipped to Fater leadership

Notebook output
Choropleth of store potential per capita by district
Per-capita store-potential metric across the 20 administrative microcode districts.How to read it: each polygon is one district; the colour scale runs from cream (low potential) to dark red (high potential). Darker means Fater is underserved there relative to population. The dark central districts are where new stores would close the gap between footprint and demand the most, and the ranking from this map was the single slide that drove the recommendation to leadership.
Recognition

Attendance certificate from Fater leadership

Signed by Fater's Sales & Digital Business Analyst Manager, Head of Data & Analytics, and Sales & Digital Data Scientist Project Manager. The certificate names the exact brief (a Neighborhood Analytics model for sales-point potential in the children's diapers category) and confirms the collaboration with the MSc Data Science programme at Federico II.

View the certificate (PDF)Issued 11 April 2024, Naples
Tech stack

Frameworks and infrastructure

MySQLSQL (CTEs, window functions)PythonGeoPandasFoliumScikit-learn

Source code on GitHub.