Session 4
Data modelling: building information systems for research
Francesco Beretta 
Beretta Taught by Francesco Beretta.
In this session
- What is a model in the context of DH research
- Conceptual modelling: producing a conceptualisation of a domain of interest
- Choosing data structures: HTML/XML, key-value, relational and graph data structures
- Two exercises
Conceptual modelling
What is the point of view of (Flanders and Jannidis 2018): what do the authors state, what do they omit?
The essence of the ontoing digital transformation is converting available information into digital data using a model. This has a huge potential in view of developing new knowledge production methods in HSS.
But what is a model? In science, a model is defined as a simplified yet equivalent representation of an original phenomenon, highlighting specific characteristics for a particular purpose. Models are used to explain, analyse or predict complex phenomena. Cf. the entry Model in Wikipedia and compare the versions in different languages.
In humanities, creating a conceptual model of a domain of interest is an essential step that provides a clear understanding of the research endeavour and should take place before any choice of methodology and technology used in data producion.
Cf. these slides that will be commented during the course.
Data structures kinds:
You can implement the conceptual model of a domain of interest using any of these data structures and technologies; each has its own specific advantages and limitations. We will explore the possibilities offered by relational databases, with a particular focus on the SQLite technology.
- HTML / XML / TEI: markup, semantic trees [this will be presented and discussed in sessions 6 and 7]
- JSON: key-value structure, lists, collections
- SQL Databases: tables, columns and relations
- Graph Databases: nodes, properties and arcs
- Labeled-property graphs: neo4j; PostgreSQL with the Apache AGE extension and in future versions implementing directly the GQL standard
- RDF [this will be presented and discussed in session 5]
Exercice 1. Inspect the chinook database: understand the data model, inspect the content
You could install the SQLite Viewer extension for Chrome or for Egde, or the SQLite Manager extension for Firefox, then dowload the database from the GitHub repo and explore it in your browser.
Reflection: diversity of domains in the humanities
Different disciplines model different things, you can build conceptual models for any of them, before you choose a methodology or technology:
- Texts — corpora, text structure, language, topics, information extraction
- Prosopography — people, roles, relations, uncertain identity across sources
- Materiality — objects, substances, condition, provenance, physical description
Exercise 2: Conceptualising some aspects of prosopography
Begin with a research question or type of analysis that you would undertake:
- map the birthplaces of your population and their development over time.
- analyse teacher-student relationships using social network analysis.
- analyse networks of organisations through memberships at different times.
How would you model the information needed to create these analyses?
List the information required and sketch a conceptual model on paper.
The Lab introduces draw.io (as VSCode extension: Draw.io Integration by Henning Dieterichs), QuickDBD and DBeaver the same week.
The DH Lab turns this into practice: replacing Excel with SQLite.
Reading for Session 5
Beretta, Francesco. 2024. “Conceptualising Information Production in the Context of the SDHSS Ontology Ecosystem.” Methodos. Savoirs et Textes 24. https://doi.org/10.4000/12xqn