The Archive at a Glance
How interactive data visualizations enhance archival research
Historical archives can be dauntingly messy. The Mediceo del Principato at the State Archives of Florence is a case in point. Indeed, only a small part of it is inventoried, and many of its documents are scattered through more than 6,500 volumes for no apparent reason. To further complicate things, the archives only let you consult a limited number of volumes (or filze as they call them). In “normal” times, the limit is set to 4 filze per day. In times of pandemic, however, that number has gone down to 4 every two weeks. In the absence of detailed inventories, the archive’s considerable size forces researchers to devise strategies in order to quickly find the documents they are looking for.
Some may privilege chance, others will also try to make educated guesses on the grounds of chronology, addressees, authors, origin of the archival fonds, language, etc. Yet looking at all these variables simultaneously may reveal unexpected patterns about the archives structure and improve our conjectures. Indeed, my experience shows that, when graphed, the metadata researchers usually gather in a spreadsheet can significantly increase situational awareness in the archive.
My current research focuses on the correspondence of a 16th-century spy. His letters are spread out through hundreds of filze. They are written under different identities, to different and sometimes unexpected addressees, from different places, etc. To find filze that are more likely to contain the expected letters, I have set-up a dashboard, an interactive data visualization web application (Plotly Dash) that connects all sorts of data, including geographical and chronological information with a hierarchic diagram (sunburst) of the archival fonds. The dashboard tells me at a glance what has already been found, how much this represents, and gives me a rough idea of where I could possibly look for new letters. By clicking on specific variables, moreover, all the diagrams are updated to show specific correlations.
Perhaps more importantly, this dashboard can be repurposed as a visual index. When the critical edition of these letters will be published on-line, the dashboard will work as an alternative entry point, from where readers will be able to browse the data. For reasons of confidentiality, I can only show a redacted screenshot at the moment, but I will release the complete dashboard next year. In the meantime a prototype will be soon available. Stay tuned!