Initially, about 30,000 comments from guestbooks were obtained for the years 2016, 2017, 2018. They come from entities of Paris Museums. In addition to comments, we sometimes have a pseudonymized username, age, country of origin, language, department if it comes from France.
A loading of new data is done whenever necessary. This operation has been automated.
See the document Data Layout Guestviews (Private) for more details.
Opinion analysis tests were started (as of 3/10/2018) to estimate whether the comment appears positive, negative or neutral (or not determined).
The data collected with the API is stored on Teralab/NextCloud in the Sources shared folder, subfolder SourceGuestviews/ParisMusees. Pseudonymized data is in json files per year (2016, 2017, 2018).
A TTL subfolder contains an RDF translation of these same data so that it can be incorporated into knowledge graphs.