Open rcavaliere opened 2 years ago
Now its deployed on testing environment
@dulvui I have tested the new DC, it works perfectly. The only thing I have seen, is that in theory we could reuse some existing types (e.g. types with id 5921... 5929 in the type table), while the DC creates all new types (types 6037 up to 6068). In particular the types 6037...6044 could be reused. The only advantage to foresee new types is that we can use them to consider forecasts computed with the old model rather than with the old model. I would like to have an opinion by @Piiit on which approach we should follow
@rcavaliere I am not exactly sure what you mean with old model, but I think reusing types is a good practice. We could then still just use older measurements for some prediction models and newer measurements for newer prediction models. So we define inputs by a timepoint and not type...
@Piiit thanks for the feedback, I will ask CatchSolve to change the Data Collector, so that the previous types are reused
@dulvui now the Data Collector is OK, please put it in production
The dc is now in production, I just need some time to adjust the old datatypes (uppercase) to the new ones (lowercase) because I accidentally deployed to production with the old datatypes
The new prediction model has been implemented by Chris Mair on behalf of STA.
To dos:
220228_SpecificheIntegrazioneODH_PrevisioniParcheggiSTAv1.2.pdf