Thesis — End-to-End BI System for Altadis

From data to decision

The Project

Master’s thesis for the Master’s Degree in Business Intelligence (UNIR), built as a team: a complete BI system for Altadis (Imperial Brands) that turns scattered data into sales, logistics and inventory decisions.

🎯

Goal

Integrate multiple data sources into a single BI system that optimises decisions on sales, shipments, stock-outs and delivery routes.

🏗️

Architecture

ETL in R (tidyverse) → PostgreSQL Data Warehouse with a star schema (4 fact tables) → executive dashboards in Power BI.

📈

Results

KPIs and interactive visualisations per area, plus clustering (K-Means, PCA) and predictive models to optimise commercial and logistics decisions.

The Dashboard in Action

A video tour of the interactive Power BI report: sales, shipments, stock-outs and routes.

What the System Includes

Power BI report cover for Imperial Brands (Altadis)

Four analysis areas

  • Sales: detailed analysis by location and performance by product and size.
  • Shipments: logistics monitoring with a geographic focus to spot bottlenecks.
  • Stock-outs: missed-opportunity losses by store and time-based incident analysis.
  • Routes: delivery optimisation and service efficiency per point of sale.

Navigation designed with storytelling: from the general to the particular, with a single source of truth thanks to the data model.

R / tidyverseSQLPostgreSQLData WarehousePower BIDAXK-MeansPCAETL

Project Documentation

Browse the presentation and the full document right here, or download them as PDF (in Spanish).

The Presentation

The slides defended before the examination board.

Download the presentation (PDF)

The Full Document

The complete Master’s thesis report.

Download the thesis (PDF)

Also published on my LinkedIn profile.