The project data is stored on a MySQL database, which powers the project website. The data is drawn from a Microsoft Access database, in which the data was stored during the editorial and transcription process. The Microsoft Access database includes data from a comprehensive census of all extant manuscript material relating to Bodley (including letters which mentioned him), although only the epistolary material written in English and written by or to Bodley falling between the years 1585-97 was transcribed for the Diplomatic Correspondence project.
A preliminary rationalization of the data in the Access database was required before embarking on creating the visualizations for Joined-up Diplomacy. This was because sections of the data had not been used for the final editorial framework of the project, so had not been part of the final rationalization process prior to the project going live.
After the data had been ‘cleaned’ the project Research Assistant Jaap Geraerts used the Open Source software Gephi to produce initial visualizations. We also investigated alternative methods of visualization including Hive plots, bar charts and pie graphs. This is because we found that while the ‘hairball’ style of visualization produces results which are visually arresting and provide a useful overview of a dataset, not only were important nuances and contexts of the dataset lost, but presenting static images of network diagrams produced in Gephi was generally unsatisfactory. Due to these limitations of the visualization software, Jaap also sought alternative graphic software to improve and clarify the visualization results.
Joined-up letters
One immediate challenge of the project was to accommodate the different editorial demands of the two projects. The original Diplomatic Correspondence had included ‘correspondence’ relating to Bodley between the years 1585 (his inaugural diplomatic mission as special ambassador to the Danish Court and to the Duke of Brunswick) and 1597 (when he was finally recalled home from The Hague). ‘Correspondence’ was taken to mean ‘all epistolary output’ relating to Bodley’s embassy, which included his letters of instruction, draft memoranda, his cipher, his passport, and other materials of a diplomatic nature. With Joined-up Diplomacy this posed a problem, as some of these materials do not have explicit addressees or correspondents, or might not be directly classed as ‘letters’ when browsing through an archive (although for scholars of correspondence networks they retain their strong epistolary connection as they were materials sent with letters within a letter-packet).
Some of these letters have multiple addressees and correspondents. Although a visualization network of people requires at least 2 nodes and 1 edge to function optimally, the porous boundaries of epistolary culture and the limitations of current visualization software mean that both correspondents and/or recipients were included as separate nodes, generating multiple edges, and meaning that a single letter might have three or more nodes. When visualized, this illustrates the wide and nebulous nature of Bodley’s correspondence network.
For instance, diplomatic letters of instruction, ostensibly from Queen Elizabeth, were drawn up by her Privy Council, and in turn letters from the Privy Council were ‘written’ by numerous hands, in terms of a supervisory overview of the Principal Secretariat. A letter like the one from the Privy Council, signed by the attending councillors Sir William Cecil, Lord Burghley, Sir Thomas Sackville, Earl of Dorset, Sir John Perrot and Sir John Fortescue (BL, MS Cotton Galba D VII f.152r-v; DCB_0927) has multiple authors, although the general author of ‘Privy Council’ means it can still function within the traditional dyad of the standard node-edge-node system in the network visualization. The occasions when letters are written from two correspondents to one recipient, for example when Bodley and Thomas Wilkes wrote to the Privy Council (BL, MS Cotton Galba D VII f.239r-241v, DCB_0954), both Bodley and Wilkes are represented as individual nodes. Although there was an option to combine the multiple author sources into one value, our decision to maintain separate author values means that three nodes and two edges are created for one letter with two authors and one recipient. The project team considered this to be more powerfully representative of the relationships sustained by the epistolary communication rather than the letters themselves.
Similarly, Bodley’s passport is seemingly difficult to render in epistolary terms at first glance. Yet it is essentially a letter in the name of Queen Elizabeth addressed to the officials Bodley would encounter en route to his seat on the Council of State in The Hague, requesting his safe conduct. In those areas in the Diplomatic Correspondence where we were unable to definitively state the precise recipient/author, these fields were defined as ‘unknown’. For Joined-up Diplomacy, authors or recipients missing from the database had to be inferred in order to supply a correspondent identity for the node. (Edges, after all, have to lead somewhere).
During the bidding stage for the funding for this project, data visualization was gaining a distinct following in the digital humanities community. By the time we began creating our visualizations, it was clear that there was more to visualization than the now-ubiquitous node-link diagram which (depending on scale of dataset) results in a large hairball of data. In refining the node-link diagram by setting filters, applying clustering and ordering algorithms and enlarging different areas of the large network, we can begin to manage this data in a meaningful way and extract conclusions. However, it was soon clear that we would need to seek other forms of visual expression which go beyond the standard network diagram. Bar and pie charts – while decidedly old-fashioned and conventional - are a surprisingly effective visualization method, communicating a simple message clearly and quickly. Because of the nature and context of the dataset, i.e. a letter network which has directional flow, we found that using Specification and Description Language diagrams to analyse small sections of the correspondence yielded very positive and interesting results (see our Case Studies).
This sub-project has been a fascinating study into alternative methods of analysis and data manipulation, and we are grateful to the Open Humanities Award for funding us. We would be interested to hear of other projects using different methods of visualization using similar datasets. Do get in touch!