Syllabus

Statistics and Big Data (Business Statistics and Artificial Intelligence)

Catholic University of the Sacred Heart, Rome — HEMA specialistic degree
Proff. Giuseppe Arbia · Niccolò Salvini — Academic year 2026–2027

Course objectives

The course provides an advanced understanding of how classical and modern business statistical methods integrate with the rapidly expanding toolkit of Artificial Intelligence. You will develop the quantitative skills needed for data-driven decision making, with particular attention to forecasting and classification problems.

A distinctive feature of the course is its focus on how AI — machine learning algorithms and generative AI tools — can complement statistical reasoning in business applications. Emphasis is placed on interpretation, managerial intuition and strategic decision-making.

Expected learning outcomes

By the end of the course you will be able to:

Knowledge and understanding — describe the foundations of advanced statistical and predictive modelling: regularised regression, GLMs, time-series analysis, unsupervised learning, and core machine learning algorithms.

Applying knowledge and understanding — build and evaluate advanced statistical and predictive models, employ time-series and forecasting techniques for planning decisions, and use AI and machine learning to complement classical models.

Making judgements — compare the predictive performance of machine learning models against statistical models, and assess the trade-off between predictive accuracy and interpretability in managerial decision-making.

Communication skills — detect and communicate uncertainty effectively to executives, through executive dashboards, storytelling with data, and the avoidance of misleading charts.

Learning skills — develop a full AI-enhanced analytical project addressing real business problems, leveraging generative AI (e.g. ChatGPT) for data exploration, coding, simulation and reporting.

Programme

Module 1 — weeks 1–6 · Prof. Giuseppe Arbia

Statistical and AI foundations. Review of hypothesis testing, confidence intervals, OLS and notions of prediction. Introduction to machine learning principles: bias–variance tradeoff, cross-validation.

Advanced regression modelling. Multiple and nonlinear regression. GLM extensions: logit, Poisson, survival models. Regularisation and high-dimensional methods: Ridge, LASSO, Elastic Net.

Time series analysis and forecasting. Time series components. Visualization. Smoothing methods. Trend and seasonality. AR and ARIMA models. Space–time modelling. Applications in business cycle analysis, stock price movements, macro-economic trends, credit scoring, consumer behaviour, and demand forecasting in health services.

Module 2 — weeks 7–12 · Prof. Niccolò Salvini

Unsupervised learning. Classification methods: customer segmentation, hospital and patient grouping in healthcare analytics. Principal components: identifying latent drivers of customer behaviour, discovering dimensions of brand perception.

Machine learning for business analytics. Neural networks. Decision trees, random forests, gradient boosting. Model interpretability; predictive accuracy vs interpretability in managerial decision-making.

Data visualization and communication. Executive dashboards, uncertainty communication, storytelling with data, avoiding misleading charts, human–AI collaboration in the analytics workflow.

Teaching method

Frontal lectures complemented by exercises in the computer lab using R (tidyverse, glmnet, fable, forecast) and Python (pandas, scikit-learn, statsmodels, PyMC), with the optional use of generative AI for workflow acceleration. Business case studies are discussed throughout both modules.

No individual or group assignments are required. The practice material on this site is optional — see Practice.

Assessment

The exam is written, with open and closed questions covering both theory and the coding part. You may be asked to provide results, the code leading to those results, or to write code that solves a given exercise.

Every exam is sat on a lab machine. Exams are held in the computer labs — Aula 1 and the two labs next to the canteen. The lab machines already have R and every course package installed. Personal laptops are for practising on; they are never used for the exam.

Before the first intermediate, and again before the second intermediate and the full exam, we run a session on the lab machines. Nobody should be meeting the exam environment for the first time on the day.

Two routes through the course

What you sit When
Full exam one exam, whole course any exam session
Intermediate route first intermediate, then second end of October, then within the winter session

Both routes are marked out of 30 and the pass mark is 18/30.

Rules of the intermediate route

  • The route runs in order: first intermediate, then second. You cannot sit only the second, and you cannot take them the other way round.
  • Fail the first intermediate and the route is over. From that point the only way through the course is the full exam.
  • Fail the second intermediate and you sit the full exam.
  • An intermediate mark cannot be refused and re-sat on its own. If you would rather not keep it — an 18 you are not happy with — the only alternative is the full exam. This applies to both intermediates.
  • The route must be completed within the winter session, by February. The session runs to three calls — last year mid-January, end of January and mid-February — so missing the first one does not end the route: sit the second intermediate at whichever call suits you, as long as it falls inside the session.
  • The final grade is the weighted average of the two modules: 50% Module 1, 50% Module 2.
  • In line with Faculty good practices, the score split between closed (multiple-choice) and open questions is published on Blackboard ahead of each exam session, together with detailed grading criteria.

Which exam you end up sitting

Same rules as above, as a picture. Skipping an intermediate, scoring below 18 and refusing a mark all land in the same place — the full exam.

flowchart TD
    S([Start of the course]) --> A{First intermediate<br>end of October}
    A -->|18 or more, and you keep it| B{Second intermediate<br>by February}
    A -->|anything else| F[Full exam<br>on the whole course]
    B -->|18 or more, and you keep it| D([Done<br>50% Module 1 + 50% Module 2])
    B -->|anything else| F

Registering

Do you register?
First intermediate — end of October No. There is nothing to sign up for; just be there.
Second intermediate, and the full exam Yes — the usual way, through the enrolment portal.

Prerequisites

Introductory statistics (descriptive statistics, estimation, hypothesis testing), basic regression analysis (OLS), and familiarity with R.

Working students are invited to contact Prof. Salvini for specific arrangements — he covers working students for both modules.

For thesis assignments, consult the lecturers’ Personal Pages (PPD) on docenti.unicatt.it and Blackboard.

Bibliography

See Resources for links and reading notes.

  • James, G., Witten, D., Hastie, T., Tibshirani, R. (2021). An Introduction to Statistical Learning. Springer. Free online
  • Provost, F., Fawcett, T. (2013). Data Science for Business: What You Need to Know About Data Mining and Data-Analytic Thinking. O’Reilly.
  • Shmueli, G., Polak, J. (2024). Practical Time Series Forecasting with R. Axelrod Schnall Publishers.

Additional readings and support material are made available on Blackboard.