Add hsanci to nssp #2162
Add hsanci to nssp #2162
Conversation
nssp/delphi_nssp/constants.py
Outdated
| "state", | ||
| "county", | ||
| "hhs", | ||
| "hsanci", |
There was a problem hiding this comment.
why not this?
| "hsanci", | |
| "hsa_nci", |
There was a problem hiding this comment.
acquisition will think the geo is hsa and the name of the signal is nci_actual_signal_name based on the csv file name.
There was a problem hiding this comment.
geos with underscores in them break an acquisition regex
There was a problem hiding this comment.
how about this then?
| "hsanci", | |
| "hsa-nci", |
nssp/delphi_nssp/run.py
Outdated
| elif geo == "hsanci": | ||
| df = df[["hsa_nci_id", "val", "timestamp"]] | ||
| df = df[df["hsa_nci_id"] != "All"] | ||
| df = df.groupby(["hsa_nci_id", "timestamp"])['val'].min().reset_index() |
There was a problem hiding this comment.
The data source reports at the HSA-NCI level and duplicates the same value across the constituent counties. This picks out just one of those per key.
There was a problem hiding this comment.
ah ok, for deduplicating. in that case, the .min() is misleading/confusing and could use a small explanatory comment
There was a problem hiding this comment.
Fair enough. Wrote a similar comment on the epidata-etl side.
There was a problem hiding this comment.
an alternative couldve been:
df = df.drop_duplicates()
Description
Add hsanci geo level to nssp
Ran through extensive existing unit tests for nssp, ran the indicators and looked at the csv files. Result looks good.