MazMarginalsCsv

These MAZ-level variables are used as marginals or contraints for PopulationSim. The new variables are bolded below. The table should be set up with a row for each MAZ and columns for each of these variables. The internal names are constant year-to-year and the table is produced for each forecast year. The divisions below are for clarity but shouldn't be in the table. A template of what the table should look like is at

Zone Number and HH Sum

• MAZ: integer, from data/parcels_geography.csv
• num_hh: simple sum of number of households (in any income category) in that MAZ

HH Size

The idea here is produce forecasts for each future year that depend on two things:

1. The proportion of each MAZ’s HHs that are in that size category in 2010. This data comes from bayarea_urbansim/data/maz_forecast_inputs.csv. A few MAZ don’t have share listed (#DIV/0! in the table) for 2010 because there are no HHs; for these please sub in the regional average 2010 proportion which is the variables shrs1_2010, shrs2_2010, shrs3_2010, and shs4_2010 in the bayarea_urbansim/data/region_forecast_inputs.csv.
2. The shift in the region-wide forecast for these categories. This data comes from the regional forecast for HH size which is from regional_demographic_forecast.csv We may need to discuss this but I see it as taking multiplying the 2010 proportions by the MAZ-level HH counts to get a floating point estimate of the number of HHs in each category. Then these numbers are IPFed to conform with the regional totals approximately and then integerized.

• hh_size_1: get the forecast year region-level proportion from shrs1 (this is short for "SHaRe Size category 1) in regional_demographic_forecast.csv
• hh_size_2: get the forecast year region-level proportion from shrs2 in regional_demographic_forecast.csv
• hh_size_3: get the forecast year region-level proportion from shrs3 in regional_demographic_forecast.csv
• hh_size_4_plus: get the forecast year region-level proportion from shrs4 in regional_demographic_forecast.csv

Group Quarters

The counts are taken directly from bayarea_urbansim/data/maz_forecast_inputs.csv

• gq_num_hh
• gq_type_univ
• gq_type_mil
• gq_type_othnon