~/teaching · data analytics · université catholique de lille

Data analytics for business, the course notes

One public summary per session of the Université Catholique de Lille course: the key ideas and a worked example

The slides of the course are encrypted for its students. Each session also has a public page: what it teaches, one worked example, the glossary, and the plan of the deck.

The six sessions follow one case, the order history of a marketplace, from the question a manager asks to the pitch of the findings.

The course page: sessions, dates, assessment and the encrypted slides

The course, in order

Session 1
From data to decisions: the analytics ladder, KPI trees and honest data cleaning
Frame a business question, break a goal into a KPI tree, and clean a real e-commerce dataset before you count anything: a manager's method.
Session 2
Data visualisation and storytelling: choosing the chart, fixing misleading ones, writing the finding
How the eye reads a chart, how to pick the right one, four ways charts mislead, and how to turn a dashboard into a finding a manager can act on.
Session 3
SQL essentials for managers: reading queries, checking joins and auditing AI-written SQL
Read and check SQL as a manager: keys and grain, SELECT, WHERE, GROUP BY, joins that duplicate or drop rows, and how to audit AI-written queries.
Session 4
Predictive analytics for decision-makers: what models learn, four traps, and the price of each error
What a model learns, four traps (overfitting, leakage, biased samples, false causes), and how to price a confusion matrix to choose between two models.
Session 5
Generative AI, GDPR and the EU AI Act in 2026: what an analyst and a manager need to know
Use an AI assistant on a data project and check its output, then the GDPR questions and the EU AI Act risk tiers and dates as amended in 2026.
Session 6
Pitching data insights to decision-makers: recommendation first, three pieces of evidence, honest answers
A ten-minute structure for presenting a data-backed recommendation, what convinces an executive audience, how to handle questions and the usual mistakes.