Currently, we do not use time-related features during training, It could be interesting to include them in our pipeline. They might be relevant proxies for tourism and or accidental fire ignitions for instance.
🚀 Feature
Implement a DateTransformer preprocessing object
Design Proposal
The DateTransformer should:
[ ] Inherit from sklearn BaseEstimator, TransformerMixin
[ ] Take and return a DataFrame
[ ] Return at least the following features (day of the week, week-end, week, month, quarter, semester)
Motivation
Currently, we do not use time-related features during training, It could be interesting to include them in our pipeline. They might be relevant proxies for tourism and or accidental fire ignitions for instance.
🚀 Feature
Implement a DateTransformer preprocessing object
Design Proposal
The DateTransformer should:
BaseEstimator
,TransformerMixin
DataFrame
[ ] Return at least the following features (day of the week, week-end, week, month, quarter, semester)
Let me know what you think 🌲