Data Science Atelier

The Data Science Atelier is a project-based course centred on statistical consulting and applied data analysis. The course combines methodological instruction with the practical experience of solving a real-world data science problem for an external client.

During the first four weeks of the semester, students are introduced to four case studies drawn from different application domains. For each case study, students are presented with a substantive research question and an associated dataset. Through guided analyses, they learn how to perform exploratory data analysis, assess data quality, formulate appropriate modelling strategies, and interpret results. The analyses focus primarily on regression methods, generalized linear models, mixed-effects models, and multivariate statistical techniques.

Following this introductory phase, students form consulting teams of approximately four members and undertake a substantial data science project. Each team works with a client who presents a practical problem requiring quantitative analysis. The project follows the complete consulting workflow:

1. Meeting with the client to understand the practical problem and project objectives.
2. Translating the practical problem into a precise statistical or mathematical question.
3. Assessing data quality through exploratory analysis and data validation.
4. Performing exploratory analyses to gain insight into the underlying phenomena.
5. Conducting formal statistical analyses using appropriate modelling techniques.
6. Formulating evidence-based answers to the original question.
7. Critically evaluating modelling assumptions, limitations, uncertainties, and potential sources of bias.

At the end of the semester, each team presents its findings to the client and submits a professional consulting report documenting the methodology, results, conclusions, and recommendations.

The Data Science Atelier provides students with hands-on experience in applied data science through the analysis of real-world problems. The course aims to bridge the gap between statistical methodology and practical application by exposing students to the complete data-analytic workflow, from problem formulation to the communication of results.

Upon successful completion of the course, students will be able to:

• Translate substantive research and business questions into well-defined statistical and data-scientific problems.
• Conduct exploratory data analyses to assess data quality, identify patterns, and generate hypotheses.
• Apply appropriate statistical methodologies, including regression models, generalized linear models, mixed-effects models, and multivariate techniques.
• Critically evaluate the assumptions, limitations, and potential pitfalls of statistical analyses.
• Communicate analytical findings effectively to both technical and non-technical audiences.
• Work collaboratively in teams and interact professionally with clients and domain experts.
• Produce clear and reproducible reports that support evidence-based decision making.

• Industria, innovazione e infrastrutture
• Pace, giustizia e istituzioni forti

In presenza

The course is based on active and experiential learning. Students learn through a combination of case-study discussions, guided practical analyses, teamwork, and client interaction.

Teaching activities include:

• Instructor-led presentations of real-world case studies.
• Hands-on data analysis sessions using statistical software.
• Group discussions on methodological choices and interpretation of results.
• Client meetings and project supervision sessions.
• Team-based project work.
• Oral presentations and peer discussion of project outcomes.

The emphasis throughout the course is on applying statistical and data-scientific methods to realistic problems, developing professional consulting skills, and learning how to communicate analytical results effectively to stakeholders with varying levels of technical expertise.

1. case study presentation
2. final consultation report