RINA - Démonstrateur d’utilisation de l’intelligence artificielle pour la gestion opérationnelle des risques naturels gravitaires géologiques | FEREC 2020

Chapter 1 Réunion Rockfall data analysis 1.1 Data characteristics All available data are recorded in several tables that could be used in a relational database, see Table 1.1. As it is simpler for single user and access, we just use pandas library to operate on these files ".csv". We list here available information in spatial and/or temporal domain, on RN1 road. Whole dataset is physically split in 2 periods , according to the major protection works in 2008 . 1. events.csv ( Temporal and Spatial ) contains each rockfall with its date, its position along RN1 road called PR for "Point de Repère" (in km) and its size (weight in kg, this variable is not complete) . This table count 902 events from January 2000 to May 2018, 2. rain.csv ( Temporal ) contains daily quantity of rain in mm from January 2000 to December 2019, more precisely, it is the maximum value between the 3 stations along the RN1 road(high correlation across this 3 geographical points) 3. zone_protect.csv ( Spatial ) which associate at each slice of the road, [ pr start ; pr end [ a binary value : 1 if there is a protection over the road and 0 otherwise. 4. zone_geo.csv ( Spatial ) which associate at each slice of the road [ pr 0 start ; pr 0 end [ the type of geological layer • Zone 0 : lower layer predominance • Zone 1 : upper layer predominance • Zone 2 : intermediate layer present 5. standarized_area.csv ( Spatial ) of the cliff profile every 10 m 6. SCI.csv ( Spatial ): a Standardized Concavity Index of the cliff profile every 10 m 7