URB-409 / 4 crédits

Enseignant: Di Lenardo Isabella

Langue: Anglais


Summary

This course explores how AI and LLMs can be used to analyze historical urban data. Students study city evolution (1700-now) through hands-on projects, producing a web interface and a visual booklet based on sources like maps, directories, and cadastral records.

Content

Course Description

This course explores how recent developments in artificial intelligence, particularly large language models (LLMs), that can be used to investigate the long-term historical evolution of cities. Through a combination of theoretical grounding and hands-on experimentation, students will learn how to critically use AI tools to analyze historical urban data.

The course is situated within the broader framework of urban systems, where the city is understood as the product of multiple interacting variables such as social, spatial, economic, and political that evolve across time and space. Urban systems analysis emphasizes the dynamic and layered nature of urban change: how housing, mobility, commerce, infrastructure, and land use co-evolve, and how such transformations can be reconstructed from historical sources. The course encourages students to think of cities not only as physical objects, but also as temporal systems shaped by flows, processes, and memory.

Students will work with historical datasets derived from city directories, cadastral sources, civil registries, and historical maps, covering the period 1700-Now. In some selected cases, datasets extend over even longer timeframes. The cities studied include Lausanne, Paris, Venice, Amsterdam, London, New York, Geneva, and selected U.S. cities. These sources allow for diverse questions related to urban morphology, socio-economic change, demography, commercial activity, and mobility.

The first part of the course introduces foundational concepts in both urban history and AI methods, with particular attention to the functioning and epistemological implications of LLMs. The second part is project-based: students will work on their own research question using LLMs and/or statistical tools to analyze the datasets provided.

Each week combines:

  • Theoretical framing (2h): urban history concepts, source criticism, and technical background on LLMs and AI methods

  • Practical session (2h): guided exercises using Python, NLP/statistical tools, dataset exploration, and project prototyping

Course Structure

  • Weeks 1-4

    • Introduction to urban history and historical sources

    • Technical overview of LLMs and their architecture

    • Epistemological discussions on the use of AI in the humanities

  • Weeks 5-6

    • Introduction to datasets and toolkits

    • Group formation and project design

  • Weeks 7-13

    • Project development with technical guidance

    • Peer feedback and interim presentations

  • Week 14

    • Final presentations and critical discussion

       

Expected Outcomes

By the end of the course, each group of students will produce:

1. A Web-Based Exploratory Interface

An online prototype that showcases the historical analysis developed during the project. It may include:

  • Interactive visualizations (maps, timelines, network graphs)

  • Interfaces for navigating or querying historical data

  • Comparative views between cities or time periods

This interface should demonstrate the students' ability to communicate a research narrative using both AI-generated insights and critical historical interpretation.

2. A Project Booklet / Infolio

A printable portfolio (PDF or printed) that documents:

  • The research question and urban context

  • The dataset(s) used and historical sources involved

  • The methods applied (LLM prompting, data cleaning, statistical analysis)

  • A critical reflection on the results and methodological choices

  • A visual narrative of the process (sketches, diagrams, maps, screenshots)

Keywords

Urban history; Historical data; Digital humanities; Large Language Models (LLMs); Artificial intelligence; Prompt engineering; Urban transformation; Spatial history; Historical GIS; Cadastral sources; Urban morphology; Data visualization; Natural Language Processing (NLP); Historical cartography; Computational history; Urban systems; Data-driven storytelling; Architectural heritage; Epistemology of digital tools; Human-AI collaboration

Learning Prerequisites

Important concepts to start the course

  • Interest in urban history, spatial analysis, or the digital humanities

  • Very welcome but not mandatory : Basic Python programming knowledge (e.g., working with Jupyter notebooks, pandas)

Learning Outcomes

By the end of the course, the student must be able to:

  • Design a small-scale research project in digital urban history
  • Apply tools to structured and unstructured historical data
  • Assess / Evaluate the results and the process from a historical and computational point of view
  • Explain the logic and technical foundations of LLMs and AI tools

Transversal skills

  • Plan and carry out activities in a way which makes optimal use of available time and other resources.
  • Use a work methodology appropriate to the task.
  • Chair a meeting to achieve a particular agenda, maximising participation.

Teaching methods

  • Weekly sessions combining lectures (2h) and practical exercises (2h)

  • Case-based learning through historical urban datasets

  • Hands-on experimentation with AI tools (LLMs, NLP libraries, basic statistical analysis)

  • Group-based project development with iterative feedback

  • Short in-class activities (e.g. dataset annotation, prompt testing, source critique)

  • Final student presentations and collective discussion

Expected student activities

Active participation in lectures and practical sessions
Critical reading of short texts on urban history and AI epistemology
Hands-on coding and experimentation (with guidance)
Exploration and cleaning of historical datasets
Definition and development of a project around a specific research question
Production of a web interface and a booklet documenting the project
Oral presentation of results and critical reflections in the final session

 

Assessment methods

Participation and exercises: 30%

Regular engagement with course sessions, weekly tasks, and peer feedback

Final project (team-based): 70%

Web-based exploratory prototype (code and design)

Printed or digital project booklet (PDF)
Final oral presentation and discussion

 

Resources

Moodle Link

Dans les plans d'études

  • Semestre: Printemps
  • Forme de l'examen: Pendant le semestre (session d'été)
  • Matière examinée: AI for urban history
  • Cours: 3 Heure(s) hebdo x 14 semaines
  • Exercices: 1 Heure(s) hebdo x 14 semaines
  • Type: optionnel
  • Semestre: Printemps
  • Forme de l'examen: Pendant le semestre (session d'été)
  • Matière examinée: AI for urban history
  • Cours: 3 Heure(s) hebdo x 14 semaines
  • Exercices: 1 Heure(s) hebdo x 14 semaines
  • Type: optionnel
  • Semestre: Printemps
  • Forme de l'examen: Pendant le semestre (session d'été)
  • Matière examinée: AI for urban history
  • Cours: 3 Heure(s) hebdo x 14 semaines
  • Exercices: 1 Heure(s) hebdo x 14 semaines
  • Type: optionnel

Semaine de référence

Cours connexes

Résultats de graphsearch.epfl.ch.