SashaWeinstein
commented
2 years ago To summarize my thoughts:
Functionality changes
- [x] Column order for race crosstabs needs to be fixed. Race crosstabs should go in alphabetical order and all races for a single category should be grouped together
Design Changes/Cleaning
- [x] Get rid of repeated docstring at top of all six indicator modules
- [x] regex for indicator names can be constructed inside load_clean_pop_data
- [x] several helper functions and
reorder_mapperconstant can be re-named to be more specific to category - [x] call
get_all_boroughsinstead of hard-coding boroughs - [x] if/else block to return correct geographies can be abstracted into helper function
td928
Overview
Six remaining indicattors (#41 #39 #40 #44 #199 #53) from the March 4th POP data. At first glance a massive but hopefully become actually quite managable with explanations.
Methodology
It is going to be similar to previous work to ingest population data with similar way to handle the columns. The work is broken out to smaller pieces of scripts that I thought make sense. There are two median indicators homevalue_median.py and rent_median.py which are single datapoint. Indicators share the same denominators are wrapped under the same functions units_overcrowd.py and units_housing_tenure.py and households_rent_burden.py each has three data points.
Review
The heavy lifting for this work is probably reviewing the final output but there are some patterns I want to share here.
For sanity check on those results, I did some math on the number of columns and they are roughly the following logics:
three datapoint sets (housing tenure, overcrowd, rent burden)
(4 race + 1 total pop) 3 data points 5 measurements * 2 years = 150
median single datapoint (homevalue and rent)
(4race + 1 total pop) 2 years 3 measurements = 30 columns
Column and Token Convention
the new output files should follow all the new column and token conventions that were established.