In this research seminar, we will discuss the ongoing extraction of data from the administrative archives of Dudelange, relying on an automated transcription, with layout detection and the automated extraction of temporal and spatial information from these sources to create historical maps.
The city of Dudelange has entered a partnership with the C²DH, for which they shared their residential cards. These administrative forms contain information on the population movements in and out of the city. These are being scanned and transcribed.
For the transcription, the project uses Transkribus to restitute not only the textual content of these administrative forms but also their structure, e.g., for each field, a tag is attached. Secondly, for the blocs of texts, a script to detect place and time information is being developed internally. All these steps should lead up to the structuration before import of the data to be visualised on a map and open these rich collections for further analysis, on the history of migration, housing, public health etc.
We invite all interested persons to join the discussion on how to improve this pipeline and share other potential uses for the extracted material.
Wednesday, 18 October 2023
14.00 - 15.00
C²DH Open Space