Project Documentation

The "Revolussion" of Renzo Renzi
Information Science and Cultural Heritage – University of Bologna (2024–2025)
Instructors: Marilena Daquino and Francesca Tomasi

1. Idea and Rationale

The "Revolussion" of Renzo Renzi is a Digital Humanities and Linked Open Data project dedicated to the archival, bibliographic, and audiovisual ecosystem built by Renzo Renzi (1919–2004). Renzi was a critic, filmmaker, researcher, and one of the foundational figures of the Cineteca di Bologna, where he curated and shaped film culture for decades.

His work produced an analogue network of interconnected materials — books, drawings, photographs, interviews, film documents — that anticipates the logic of Linked Open Data. Long before LOD became a methodological paradigm, Renzi was already constructing a relational archive where each object illuminated the others through themes, stories, production histories, and personal memories.

The Renzo Renzi Collection at the Cineteca di Bologna is the institutional foundation of this project. Our goal is to model, encode, and publish a selection of these materials using cultural heritage standards, TEI/XML, RDF, and a public web interface, transforming analogue archival materials into a structured, interoperable digital dataset aligned with LODLAM best practices.

2. Selection of Items

The project selects 15 heterogeneous items from the Renzo Renzi Collection and related holdings at the Cineteca di Bologna. The selection includes archival documents, bibliographic records, photographs, drawings, moving images, sound recordings, and one full-text item encoded in TEI.

3. Metadata Analysis

For each item, we identified the descriptive standard adopted by the holding institution and used it as the foundation for our analysis. This step ensured that our modelling respected institutional practices while enabling interoperability across heterogeneous materials.

Bibliographic standards

• ISBD(G) — general bibliographic description
• ISBD(NBM) — non-book materials (film, soundtrack, documentary)
• SBN-MARC — Italian national bibliography

Archival standards

• ISAD(G) — general archival description

Visual / museum standards

• ICCD Scheda F — photographs
• ICCD Scheda OA — drawings and caricatures

Audiovisual standards

• FIAF Cataloguing Rules — film-related materials

Encoding formats

• MODS — bibliographic items, sound recordings, films, documentary, video interview
• EAD — archival records
• VRA Core — photographs, drawings, caricatures
• Schema.org (JSON-LD) — institutional entity

4. Theoretical Model

The theoretical model describes each item in natural language, starting from the institutional description and enriching it with additional information such as authority control, subject classification, and contextual relationships.

For each object the analysis identifies:

• Intrinsic features: title, date, format, technique, language, extent…
• Roles and agents: creators, performers, contributors, editors, directors…
• Relationships: about, depicts, documents, is part of, was produced during…
• Institutional context: holding institution, collection, physical location, shelf mark…
• Authority control: VIAF, Wikidata, ISIL, ISNI…

This analysis maps the domain narrative connecting:

• People: Renzo Renzi, Federico Fellini, Giulietta Masina, Nino Rota, Anthony Quinn…
• Works: film, documentary, book, soundtrack, screenplay, video interview…
• Images: photographs, drawings, caricatures…
• Events: film premiere at Cinema Fulgor, documentary production along the Po, military trial…
• Institutions: Cineteca di Bologna, Biblioteca Renzo Renzi…
• Places: Bologna, Po River Delta, Cinema Fulgor…

The theoretical model is represented as an interactive diagram on the project website.

5. Conceptual Model

The conceptual model formally represents the theoretical model by reusing existing schemas, vocabularies, and ontologies — no new ontology was created. CIDOC CRM and RiC‑O were consulted as conceptual references, while Dublin Core Terms, Schema.org, FOAF, SKOS, and OWL were actively reused in the RDF dataset.

Ontologies and vocabularies reused

• Dublin Core Terms (DCTerms) — general metadata (creator, title, date, subject, publisher…)
• Schema.org — creative works, agents, places, events
• FOAF — agents and depictions
• SKOS — subject classification and authority control
• OWL — owl:sameAs for linking project entities to external authority files
• CIDOC CRM — conceptual reference model for cultural heritage
• RiC‑O — archival conceptual reference

Core classes

• schema:CreativeWork, schema:ImageObject, schema:VideoObject, schema:MusicRecording
• schema:Person, schema:Organization, schema:Place, schema:Event

Core properties

• dcterms:creator, dcterms:contributor, dcterms:subject, dcterms:date, dcterms:publisher
• schema:about, schema:locationCreated, schema:hasPart
• foaf:depicts
• owl:sameAs — linking project entities to VIAF, Wikidata, GeoNames, and OPAC SBN

The conceptual model is represented as an interactive diagram on the project website.

Authority URIs used

• Renzo Renzi — VIAF: 40486517 · Wikidata: Q56179169
• Federico Fellini — VIAF: 76315386
• Cineteca di Bologna — VIAF: 124960346
• Bologna — Wikidata: Q1891 · GeoNames: 3181928
• Cinema Fulgor — Wikidata: Q36839368

6. CSV Files

A dedicated CSV file was created for each of the 15 items, structuring metadata according to the relevant institutional standard and placing items in dialogue through shared entities and relationships.

Two global files handle the semantic layer:

• rrr_entities.csv — all domain entities (people, places, works, institutions…) with identifiers and authority URIs
• rrr_triples.csv — explicit relationships between entities, expressed as subject–predicate–object triples

All local identifiers use the shared prefix rrr: (namespace: https://github.com/CineFiles25/TheRevolussionOfRenzoRenzi/).

The CSV files are the starting point for the RDF generation. All files are available in the CSV directory on GitHub.

7. XML/TEI Document

One item — Sequence I of the screenplay of La Strada, taken from the book Il primo Fellini — was encoded as a full-text document using TEI P5.

The TEI file (tei_xslt/lastrada.xml) includes:

• a complete <teiHeader> with bibliographic and archival metadata (title, author, editor, publisher, date, language)
• a <particDesc> listing all named characters with their actors linked to VIAF authority records
• a <settingDesc> with all places appearing in the sequence
• semantic tagging of dialogues, stage directions, characters, and places within the text body
• logical structuring of scenes into labelled segments

→ View TEI/XML source

8. XML to HTML Transformation

The TEI file is transformed into a web-publishable HTML edition through an XSLT pipeline:

• The stylesheet tei_xslt/tei2html_lastrada.xsl defines the transformation rules from TEI elements to HTML.
• The Python script scripts/xml_to_html.py applies the stylesheet using the lxml library (etree.XSLT), writing the result to html/lastrada.html.

→ Open the TEI-based HTML edition of La Strada

9. XML/TEI to RDF Transformation

The TEI file is also transformed into RDF using a dedicated Python script:

• scripts/xml_to_rdf.py uses xml.etree.ElementTree to parse the TEI/XML source and extract structured metadata: title, author, editor, publisher, date, language, characters, and places.
• Extracted data is mapped to RDF triples using RDFLib, reusing Schema.org, Dublin Core Terms, FOAF, and CIDOC-CRM vocabularies.
• The output is serialized as a Turtle file (tei_xslt/lastrada_screenplay.ttl) and as RDF/XML (tei_xslt/lastrada_screenplay.rdf).

10. RDF Dataset (CSV → RDF via Python)

The full RDF dataset is produced as a set of modular Turtle files (ttl/*.ttl), one per cultural heritage item. Dedicated Python scripts in scripts/ each:

1. Read the item-specific CSV file.
2. Map the CSV fields to RDF triples according to the conceptual model, using RDFLib.
3. Serialize the output as an individual Turtle file in the ttl/ directory.

All item scripts declare the following shared namespaces: Dublin Core Terms, Dublin Core Elements, Schema.org, FOAF, CIDOC-CRM, OWL, and SKOS.

Authority linking pipeline

External authority URIs (VIAF, Wikidata, GeoNames, OPAC SBN) are centralised in csv/rrr_entities.csv, in the sameAs column. Rather than distributing authority references across individual item scripts, the project adopts a single-point-of-truth approach: merging.py reads rrr_entities.csv at merge time and automatically generates owl:sameAs triples for every entity that carries an external URI. This ensures that authority links are consistent, maintainable, and decoupled from item-level scripting.

Merging

The script scripts/merging.py:

1. Parses all individual Turtle files in ttl/.
2. Reads csv/rrr_entities.csv and injects owl:sameAs triples for all entities with an external authority URI.
3. Serializes the complete unified graph as ttl/full_dataset.ttl.

The dataset integrates:

• creative works (books, films, drawings, photographs, sound recordings…)
• agents and institutions (Renzi, Fellini, Masina, Rota, Cineteca di Bologna…)
• events (premiere, documentary production, film-related activities…)
• owl:sameAs links to external authority files (VIAF, Wikidata, GeoNames, OPAC SBN)
• inter-item relationships expressed as RDF triples

The dataset is modular by design: files can be loaded as separate named graphs or merged into a single RDF graph for SPARQL querying.

11. Website

The full project is published as a GitHub Pages website at:
https://cinefiles25.github.io/TheRevolussionOfRenzoRenzi/

The site includes: project overview, item list with metadata, conceptual and theoretical graphs, TEI-based HTML edition, RDF dataset downloads, photo gallery, team, and full documentation.

12. Team

The project was developed within the course Information Science and Cultural Heritage (a.y. 2024–2025), University of Bologna – DHDK.

• Laura Bortoli — laura.bortoli@studio.unibo.it — GitHub: lauraaa13
• Claudia Romanello — claudia.romanello@studio.unibo.it — GitHub: claudiarom
• Qinghao Chen — qinghao.chen@studio.unibo.it — GitHub: River-Qinghao

All project members contributed collaboratively to metadata extraction, TEI encoding, RDF modelling, Python scripting, and interface development.