MODELLING AND VISUALIZING TEXTUAL DATA

The course Modelling and Visualizing Textual Data is part of the Master’s Degree Programme in Digital and Public Humanities and is connected with the activities of the Venice Centre for Digital and Public Humanities (VeDPH) in the Department of Humanities. The course introduces students to the main theoretical and practical issues involved in modelling and visualizing textual data, with particular attention to literary texts. Working with a public-domain corpus centred on Joseph Conrad and "Heart of Darkness", students will examine how a text can be described, encoded, annotated, transformed into data, analysed quantitatively, and explored through computational methods. Particular attention will be paid to the relationship between explicit models, statistical models, and artificial intelligence systems.

By the end of the course, students will have acquired a critical understanding of the main ways in which literary texts can be modelled in a digital environment. They will be able to design and document simple data models, distinguish between different forms of textual representation (metadata, tables, markup, semantic annotations, networks, corpora, computational outputs), and assess the suitability of each in relation to the aim of the analysis. Practical activities will allow students to experiment with tools for textual analysis and visualization, introduced step by step and without assuming previous programming experience. The course will also enable students to discuss critically the results produced by computational methods and AI systems, recognising their potential, limits, errors, opacity, and bias. Particular attention will be devoted to the documentation of the work carried out and to the reproducibility of the analyses.

The course is placed in the second year of the Master’s Degree Programme in Digital and Public Humanities and builds on some concepts and methods introduced earlier in the programme, especially XML/TEI, data formats and structures such as CSV and JSON, and the Jupyter notebook environment. Familiarity with these elements is useful, but it is not a strict requirement: students coming from other academic backgrounds will be able to catch up through introductory materials and guided activities. Advanced programming skills are not required. Students are expected to interpret the results of the analyses critically and to document their modelling choices.

The course introduces the concept of model and the practice of modelling and visualizing literary textual data in the Digital Humanities. It follows the transition from explicit and directly documentable models to statistical, machine-learning-based, and generative models, in which procedures become progressively less transparent and harder to inspect. Starting from this trajectory, the classes invite students to ask what interpretative and technical choices are embedded in different models, what they make visible, what they simplify, and where they may fail. The main text will be Joseph Conrad’s "Heart of Darkness", which students are required to read in full. The text will be accompanied by a small public-domain corpus of works by Conrad.

The course will cover: data and metadata, dataset design, and the use of tabular formats such as CSV; formal models and digital scholarly editing, with particular attention to XML/TEI; corpus construction and analysis; distant reading and quantitative analysis; semantic annotation, RDF, ontologies, and Linked Data; the modelling of places, time, and entities; guided use of Jupyter notebooks for text processing, named entity recognition, embeddings, classification, and topic modelling. The final part of the course will be devoted to artificial intelligence as a transformation of modelling practices: automatic generation, opacity, bias, hallucination, evaluation of results, documentation, and scholarly responsibility. Recurring topics will include operationalization, the modelling of spatial and temporal data, data structures, ontologies, and the construction of reproducible workflows. Methods and procedures will always be presented in relation to interpretative and methodological problems.

The following topics will be addressed in particular:

* the concept of model, modelling, and operationalization;
* data, metadata, datasets, and tabular structures;
* data formats and structures, with attention to CSV and JSON;
* markup, XML/TEI, and the modelling of digital scholarly editions;
* corpus construction and analysis;
* distant reading, quantitative analysis, and the visualization of textual data;
* the modelling of spatial and temporal data;
* semantic annotation, RDF, ontologies, and Linked Data;
* guided use of Jupyter notebooks for text processing, named entity recognition, embeddings, classification, and topic modelling;
* critical use of artificial intelligence for textual analysis, with attention to automatic generation, opacity, bias, hallucination, evaluation of results, and documentation;
construction of reproducible workflows based on public-domain texts and data.

Students are required to complete exercises in modelling and textual analysis and to document the choices made.

Primary literary text:
- Joseph Conrad, Heart of Darkness (published in volume form in 1902). Students are required to read the whole text. They may use any complete English edition of the work, including editions accessed through digital libraries and digital lending services such as Open Library. For computational activities, a public-domain edition will be made available through Moodle. During the classes, the lecturer will refer to a critical edition of the work, in particular the Norton Critical Edition edited by Paul B. Armstrong (5th ed., 2016), as a teaching reference for textual, historical, and interpretative commentary.

Theoretical and methodological reference texts:
- A. Ciula, Ø. Eide, C. Marras, P. Sahle, "Modelling Between Digital and Humanities. Thinking in Practice", Open Book Publishers 2023. DOI: https://doi.org/10.11647/OBP.0369
Primary
- F. Moretti, "Operationalizing": Or, the function of measurement in modern literary theory", Pamphlet 6, December 2013, available through Stanford Literary Lab: https://litlab.stanford.edu/pamphlets/

Further readings on modelling, corpus analysis, Linked Data, artificial intelligence, and data documentation will be made available through Moodle during the course. Materials may include articles, book chapters, notebooks, sample datasets, and public-domain digital resources.

Assessment takes the form of an oral exam, accompanied by the discussion of a compulsory modelling portfolio for all students. The portfolio brings together a limited number of exercises and artefacts produced from the course materials, which will be available on Moodle and can also be completed independently by students who do not attend classes. The exercises may include, for example, a metadata table, an encoding or annotation exercise, a short quantitative analysis, a notebook-based exercise, a critical evaluation of a result generated by an AI system, and a documentation sheet for the model or dataset used.
The portfolio does not require advanced programming skills. It will be assessed mainly on the clarity of its structure, the coherence between method and material analysed, the quality of documentation, the ability to interpret results critically, and attention to the limits of the procedures used.
During the oral exam, students will discuss the main concepts of the course, the assigned texts and materials, and the work included in the portfolio. Students may also present an optional extended project, devoted to a specific issue in modelling or textual analysis.

oral

The instructor is responsible for ensuring the authenticity and originality of all examinations and coursework. In cases of suspected academic misconduct, an additional on-site assessment may be required during the exams, which may differ from the standard format.

The final grade will take into account the oral exam, the modelling portfolio, and any optional extended project. Grades will be assigned according to the following criteria:

18-22: basic knowledge of the course contents; limited but sufficient ability to describe methods, materials, and results; portfolio complete in its minimum required elements, but with partial documentation or underdeveloped interpretation.
23-26: good knowledge of the main concepts and methods; ability to discuss the work carried out correctly; well-organised portfolio, with identifiable methodological choices and a satisfactory interpretation of results.
27-29: solid and articulated knowledge of the course contents; good ability to connect theory, methods, and case studies; well-documented portfolio, with coherent analyses and critical attention to the limits of the procedures used.
30-30 cum laude: complete and in-depth knowledge of the course topics; autonomous and critical discussion; very well-structured portfolio, with accurate documentation, methodological awareness, convincing interpretation of results, and the ability to reflect on the theoretical and scholarly issues involved in modelling.

The course combines lectures, seminar-style discussion, case-study analysis, and guided exercises. Lectures will introduce the main theoretical and methodological concepts involved in the modelling of textual data, while practical activities will focus on the construction, analysis, visualization, and documentation of simple models applied to literary texts.

The exercises will use materials made available on Moodle, a small working corpus, tools for textual analysis, and Jupyter notebooks. Computational activities will be introduced progressively and supported by instructions, examples, and additional resources. Particular attention will be devoted to the documentation of modelling choices, the reproducibility of workflows, and the critical discussion of the results obtained.

Course materials, portfolio instructions, and supporting resources will be available on Moodle, so that students who do not attend classes can also follow the course structure and prepare for the final assessment.

All course materials, portfolio instructions, working texts, notebooks, and any supplementary readings will be made available through Moodle. Students who do not attend classes are invited to consult the platform regularly and to follow the instructions provided for the preparation of the portfolio and the oral exam. Any use of artificial intelligence tools in course activities must be declared and documented. These tools may be the object of critical analysis, but they do not replace the interpretative, methodological, and documentary work required.

Accessibility, Disability and Inclusion.
Accommodation and support services for students with disabilities and students with specific learning impairments:
Ca’ Foscari abides by Italian Law (Law 17/1999; Law 170/2010) regarding support services and accommodation available to students with disabilities. This includes students with mobility, visual, hearing and other disabilities (Law 17/1999), and specific learning impairments (Law 170/2010). If you have a disability or impairment that requires accommodations (i.e., alternate testing, readers, note takers or interpreters) please contact the Disability and Accessibility Offices in Student Services: disabilita@unive.it.

This subject deals with topics related to the macro-area "Human capital, health, education" and contributes to the achievement of one or more goals of U. N. Agenda for Sustainable Development