Visualizations of historical spatial data as tools of exploration and education
During the Human-Computer Interaction International Conference of 2022 (HCII2022), I delivered a paper on the value of (geo)visualisations as tools of exploration and education in the humanities. The extended version of the paper has been published as Visualizations of Historical Spatial Data as Tools of Exploration and Education in the “HCII 2022 – Late Breaking Work – Papers” volume of Springer’s Lecture Notes in Computer Science (LNCS) series. The paper includes several simple maps created by MA students for the DigiKAR geo-humanities project.
PDF download (for users with an institutional subscription): Visualizations of Historical Spatial Data as Tools of Exploration and Education
In this blog post, I am sharing the manuscript for my original conference presentation as well as my slides.
Visualizations of historical spatial data as tools of exploration and education – examples from the DigiKAR project:
Based on pilot studies conducted in the DigiKAR geo-humanities project, which analyzes spatial relations in the Holy Roman Empire of the seventeenth and eighteenth centuries, my presentation highlights the value of iteratively and experimentally visualizing spaces in historical research. Our experiences are meant to inspire other humanities projects, especially in historical subject areas, to experiment with visualizations even on a technically basic level, and even when sources are incomplete, fragmented or uncertain.
The project Digitale Kartenwerkstatt Altes Reich (abbreviated to DigiKAR and meaning Digital Map Lap Holy Roman Empire) is an interdisciplinary and international project led by the Institute of European History in Mainz and funded by the German Leibniz Association from 2021 to 2024. It questions established visual representations of a historical space whose legal and political complexity is only insufficiently captured through polygons and line-oriented maps with narrow thematic foci. As a fractured yet interwoven territory, the Holy Roman Empire and its individual, often overlapping or contested regions require multi-modal and multi-perspective visualizations. By proto-typically tracing human, material, and ideational border-crossings and competing spatial conceptions, the project addresses the central challenges of both traditional imperial historiography and digital historical research.
Explorational & educational geo(visualisations) in DigiKAR
One primary focus of DigiKAR is to understand data visualisations as both explorational and educational tools. Our approach to (geo)visualisations is marked by the following four statements:
- Visualisations (in the plural), not ONE visualisation!
- Visualisations not at the end of a project, but throughout!
- Visualisations not only by “experts”, but also by students!
- Visualisations on various levels of complexity!
Bringing together interdisciplinary researchers as well as heritage professionals and students, the current phase of the DigiKAR project uses maps for high-level data evaluation, data cleaning, a critical reflection on our own research biases, and as tools to highlight developments and alternative views of our data. At a later stage, more elaborate visualizations as tools of science communication will follow.
Visualisation as storytelling
Overall, we view visualisations as a storytelling device on two levels. On the one hand, maps help us to tell history from different perspectives and to make “gaps” and ambiguities visible. On the other hand, visualisations tell the story of our research process and make (changing) research questions transparent.
In my slide, I am giving an example of the narrative embedding of visualisations in our project. The DigiKAR pilot project “Early modern birth letters from Mainz“, which we completed in August 2021, combines a digital edition of multi-lingual text sources, different maps, and academic content.
What data are being visualised in DigiKAR?
The data that we visualise in DigiKAR is — similar to many other historical projects — highly fragmented and come from various sources. First of all, we work with spatial data extracted from textual (narrative) early modern sources (both in the form of archival manuscripts and printed books). Secondly, we work with spatial data extracted from early modern maps (using OCR technology for identifying place names). Thirdly, we are re-using historical spatial data originally gathered by historians in previous decades (some of which have already been digitised and made available via APIs).
The WORKFLOW in our project includes 6 main steps:
- semi-automated data collection in spreadsheets (using Python for web-scraping)
- semi-automated data normalisation and data cleaning
- geocoding with Google and Open Street Map APIs (incl. manual corrections)
- preliminary, experimental geo-visualisations in QGIS
- import of consolidated data into Metaphacts (graph database) and/or a relational GIS database
- advanced data queries and advanced visualisations created by the cartography team at IfL Leipzig
What visualisation types are used in DigiKAR?
As the DigiKAR team is diverse and committed to the further training of all members, the visualisation we use in DigiKAR range from simple, “home-made” maps to high-end animations. All visualisations, however, are important to our research process and describe our data from several perspectives:
- Maps in which point geometries are displayed in different colours indicating developments over time
- Maps in which point geometries are clustered to highlight areas of intensive (administrative) activity
- Heatmaps (highlighting “fuzziness”)
- Maps in which point geometries are linked to show people’s mobility
- Interactive, zoomable web maps which permit the (de)selection of several layers for comparison
- Maps created by professional cartographers (IfL Leipzig) using interpolation and continuous rasters (e.g. to show zones of legal conflict)
The image gives an example of a static map created for DigiKAR: In this map, lines are connecting early modern migrants’ places of birth with places where birth or apprenticeship letters were issued. These letters are the early modern versions of passports and other certificates. The spatial data behind this map came from early modern manuscript records kept at the City Archives in Mainz, Germany.
Some metadata had already been digitised, other data were collected by a team of MA students who transcribed letters from scratch. The visualisation displayed here was created by MA student Maximilian Michel and helped us understand the often impressive distances across which early modern agents would relocate for private or professional reasons.
Our most recent mapping experiments cover the professional biographies of high-ranking clerics in the Prince-Archbishopric of Mainz between 1600 and 1800. In collaboration with MA students of history, who are taught digital methods through working with us, we create both static and interactive visualisations and will make them available via our DigiKAR Projektseminar site. These maps will help us gain further insights into the administrative structures of Mainz as an early modern ecclesiastical principality.
All in all, the process of creating visualizations such as maps for historical research projects is like an iceberg. The visual output is only the uppermost part of the technical and human effort that goes into it. Also, historians cannot rely on just one tool or methodology. Sources must constantly be analyzed in comparison with others, including sources of entirely different genres. This also requires seeing abstract structures in text and reading historic maps as narratives. Visualizations we create are hardly ever final but reflect work in progress. This is why the DigiKAR project shares data tables, scripts and exemplary visualizations on GITHUB where they can be easily updated and versioned. At the same time, we want our data to be inter-operable and open for re-use, which is why we apply established data models whenever possible. Beyond the specific research questions of the DigiKAR project, we intend our data modelling and visualization efforts to serve as samples for similar humanities projects. As explained in this presentation, our maps are not primarily research output but flexible tools of exploration. Moreover, experienced historians, as well as students and interns, benefit from making maps themselves as it challenges them to name uncertainties and ambiguities in their research data.
If you want to find out more about DigiKAR, you are welcome to visit our website. Questions can be addressed to firstname.lastname@example.org. My colleague Constanze Buyken will be happy to respond.
Download all slides in PDF format: Blog_VisualizationAsExploration