This section collects the resources that Creative Data Exchange has been showcasing in our events and community activities. We are forming a library of resources that support creative uses of data, software and creative practice.
On this page
Open notes from our Community Call, 4 September 2026
Explore the CoSTAR Foresight Lab’s research on the impact of technology in the creative industries
Moments, Praxis, Paths is a film presenting 4 foresight scenarios for the creative industries in 2040 based on the CoSTAR Foresight Lab’s research. Watch the trailer below and register your interest to arrange a screening for your organisation.
Explore the CoSTAR Foresight Lab’s prototyping tools, navigate through the lab’s publications and check the lab’s research tracking the adoption of generative AI tools across film, television and videogames.
Morphing Wings
Explore this interactive project on bird morphology and wing structures. This visual and engaging resource demonstrates how data can be used to reveal patterns in nature and supports creative approaches to data exploration.
Marker: creative and collaborative web annotation and writing
Marker is a writing tool that allows you to save, annotate, organise, and share content from across the web. It can be useful for research, collaboration, knowledge sharing, and collecting inspiration from online sources.
MapReader resources and further reading
Explore the tools, datasets and research behind MapReader, an open-source computer vision framework developed for analysing historical maps and other image collections at scale.
Learn about the research project behind MapReader
- Living with Machines
A pioneering interdisciplinary research project exploring how technology transformed everyday life during the Industrial Revolution. MapReader was originally developed as part of this initiative to help researchers analyse thousands of historical maps using AI and machine learning techniques.
Explore the historical map collections
- National Library of Scotland Historic Maps
Access extensive digitised collections of Ordnance Survey maps of England, Wales and Scotland. These historical maps form one of the key datasets used in MapReader research. - Georeferenced Historical Map Viewer
Compare historic maps with contemporary geography through an interactive map explorer. A useful resource for understanding the types of materials that MapReader can analyse.
Access the software
- MapReader GitHub repository
The main open-source software repository containing the MapReader codebase. The platform supports image classification, feature detection and text extraction workflows for historical maps and other image datasets.
Documentation and getting started
- MapReader documentation
Comprehensive guides covering installation, tutorials, workflows, API references and best practices for using MapReader in research and creative applications.
Examples and tutorials
- MapReader examples repository
A collection of worked examples, Jupyter notebooks and reproducible workflows demonstrating how to use MapReader for both geospatial and non-geospatial image analysis projects.
Research publications
- Patch-based classification for historical maps
Learn about the machine-learning concepts that influenced MapReader’s approach of classifying image patches rather than relying exclusively on pixel-level segmentation. - Combining text recognition and map analysis
Research demonstrating how MapReader combines OCR and image classification techniques to discover patterns within historical map collections. - Beyond historical maps
An example of how MapReader-inspired methods can be applied in other domains, including plant phenotyping and scientific image analysis.
Computational humanities and AI resource guide
Explore a selection of projects, datasets and research tools demonstrating how computational methods can be used to study culture, history, literature and censorship. You will find materials that:
- Demonstrate how computational methods can uncover patterns hidden in large cultural collections.
- Show how AI can be used both as a research tool and as an object of critical investigation.
- Connect digital humanities, archival research, data science and creative technologies.
- Inspire new approaches to working with large-scale textual and cultural datasets.
Historical newspapers and spatial analysis
- Mapping Victorian Accidents 1 (animated visualisation)
A visualisation of accidents extracted from nineteenth-century newspapers, built using place-name detection and disambiguation to locate accident reports geographically. It shows how accidents can be analysed across both time and space using computational methods applied to historical newspaper collections. - Mapping Victorian Accidents 2 (second case-study visualisation)
A companion visualisation from the same Victorian accidents project. It demonstrates large-scale analysis of historical newspaper archives and illustrates how computational approaches can reveal patterns and narratives hidden within vast textual collections.
Censorship, queer archives and digital collections: AI, language models and cultural data
Investigate how large language models interpret, classify and moderate literary texts. Explore practical examples of prompt design, content classification and bias analysis, and reflect on the role of AI systems as cultural gatekeepers and their influence on access to information.
- Seized Books Gallery
An online gallery from the University of London’s Seized Books project. It showcases books confiscated during “Operation Tiger”, a UK customs campaign targeting imported LGBTQ+ literature, and provides access to digitised materials from the Cutbill Collection, exploring questions of censorship, sexuality, cultural value and archival preservation. - Seized Books Research Notebook (Google Colab)
A computational notebook associated with the Seized Books project, used to investigate how large language models respond to and classify passages from LGBTQ+ literature. It supports research on AI moderation, censorship, bias, moral reasoning and content classification.