Historical uncertainty in Gregory of Tours's History of the Franks (book 7)

Our goal was to create a research dataset based on geographical and chronological uncertainties in the work of Gregory of Tours's *History of the Franks* (book 7). We used and modified a topology of geographical and chronological uncertainty based on a rudimentary schema that would be universal when analysing an historical source :

Chronological :
* uncertain dating
* uncertain method of dating
* lack of dating
* precise dating

Geographical: 
* uncertain location
* general location (region, country)
* lack of location 
* precise location

After working on book 7 for a while, that schema was reworked as those 9 types of uncertainty : 

Chronological :
* uncertain_dating
* uncertain_method_dating
* event_dating_null
* precise_dating

Geographical: 
* uncertain_location
* general_location 
* event_location_null
* uncertain_method_location
* precise_location

The geographical and chronological focus makes it possible to identify where and when, in a source, the historical uncertainty is higher. 

Using python, that dataset was then automatically cleaned and enhanced with bounding box based on geo-mapping information for the entries of geographical uncertainty. Those were classified as either precise_location or general_location. 

For example, anything relating to a city general area (like  "in the Rouen area") creates a general_location bounding box encompassing the *current* geographical space occupied by the municipality of Rouen (in the format 'LongMin', 'LongMax', 'LatMin', 'LatMax' in a single column "bbox"). Anything described as a unique point in space (like "in Paris") creates a precise_location and its corresponding lat/long system of coordinates. 

This is an arbitrary way to translate slightly undefined geographical concepts of uncertainty into formal data, but at least it can be fully explained explicitly.
  

Translation used: Tours G. et alii, The history of the franks, Penguin Books Limited, 1974, https://books.google.ch/books?id=4Lx-M2RHGgoC.