Tools for testing
The tool Google Fusion at first allowed us to consolidate the data from the spreadsheet and to test various methods of representation. Thanks to this tool, it is rather easy to “play” with the data to test the pertinence of the representations.
Figure 2.2. Map of the locations visited by Jules Michelet. For a color version of the figure, see www.iste.co.uk/szoniecky/collective.zip
This map shows the locations visited by Jules Michelet as well as the frequency of these visits: the brighter and redder the location is, the more the location is visited. The interest of this map is to quickly evaluate the activity zones of this person or, conversely, the places he did not visit. Clearly, this visualization lends nothing to a Michelet expert, who has memorized his journeys and his vacations, and most of us expected to see Normandy appear often in the locations visited by Flaubert, Maupassant, Michelet, etc., but in general, it is rather the establishment of relationships in maps concerning lesser-known actors who may lead to the appearance of unexpected cross referencing. The following representation, also with Google Fusion, aims to characterize the types of relationships in association with the location of the stay.
Figure 2.3. Relationship between the types and locations of stays (1st attempt). For a color version of the figure, see www.iste.co.uk/szoniecky/collective.zip
This first visualization work makes the researchers aware of the potential inconsistencies in the data, particularly in terms of duplications, unusable information or ambiguity in the choice of categories. For example, the difference between “journey” and “stay” had to be made explicit to understand that in the former, it was a question of a voyage supposing short says in each town, while in the latter, one of real, long-term moves, in a new space allowing more than quick visits to foreign scholars, for example.
Google Fusion thus proved to be an excellent “pre-test” tool, despite the questionable esthetic of its visualizations, highlighting the weaknesses of certain categories, and allowing the quick and easy correction of the initial database, which was supposed to provide a model for the other datasets.
-  To read the Google Fusion documents, see https://goo.gl/Tm83jp.