In [1]:

    

import numpy as np, pandas as pd, os
from synthicity.utils import misc
from drcog.models import regression_model_estimation, choice_model_estimation, dataset
dset = dataset.DRCOGDataset(os.path.join(misc.data_dir(),'drcog.h5'))
np.random.seed(1)

##Variable Library
from drcog.variables import variable_library
variable_library.calculate_variables(dset)

buildings = dset.fetch('buildings')[['building_type_id','improvement_value','land_area','non_residential_sqft','parcel_id','residential_units','sqft_per_unit','stories','tax_exempt','year_built','bldg_sq_ft','unit_price_non_residential','unit_price_residential','building_sqft_per_job','non_residential_units','base_year_jobs','all_units']]
establishments = dset.fetch('establishments')
del establishments['zone_id']
del establishments['county_id']
households = dset.fetch('households')
del households['zone_id']
del households['county_id']
parcels = dset.fetch('parcels')
parcels_urbancen = dset.store.parcels_urbancen.set_index('parcel_id')
parcels['urbancenter_id'] = parcels_urbancen.urban_cen
zones = dset.fetch('zones')
pz = pd.merge(parcels.reset_index(),zones,left_on='zone_id',right_index=True,how='left')
pz = pz.set_index('parcel_id')
bpz = pd.merge(buildings,pz,left_on='parcel_id',right_index=True)

##Merge buildings and parcels
buildings = pd.merge(buildings,parcels,left_on='parcel_id',right_index=True)

##Merge households with bulidings/parcels
households = pd.merge(households,buildings,left_on='building_id',right_index=True)

##Merge establishments with bulidings/parcels
establishments = pd.merge(establishments,buildings,left_on='building_id',right_index=True)

#####Export household points
hh = households[['building_id']].reset_index()
hh['parcel_id'] = bpz.parcel_id[hh.building_id].values
hh['urbancenter_id'] = bpz.urbancenter_id[hh.building_id].values
hh['x'] = bpz.x[hh.building_id].values.astype('int64')
hh['y'] = bpz.y[hh.building_id].values.astype('int64')
hh['taz05_id'] = bpz.external_zone_id[hh.building_id].values
hh['dist_trans'] = np.minimum(bpz.dist_rail[hh.building_id].values, bpz.dist_bus[hh.building_id].values)/5280.0
big_parcels = parcels.index.values[parcels.parcel_sqft>= 435600]
big_parcel_ids_with_hh = np.unique(hh.parcel_id[np.in1d(hh.parcel_id,big_parcels)].values)
parcel_coords = dset.parcel_coords
parcel_coords.x = parcel_coords.x.astype('int64')
parcel_coords.y = parcel_coords.y.astype('int64')
for parcel_id in big_parcel_ids_with_hh:
    idx_hh_on_parcel = np.in1d(hh.parcel_id,[parcel_id,])
    coords = parcel_coords[parcel_coords.parcel_id==parcel_id]
    idx_coord = np.random.choice(coords.index,size=idx_hh_on_parcel.sum(),replace=True)
    x = coords.x.loc[idx_coord].values
    y = coords.y.loc[idx_coord].values
    hh.x[idx_hh_on_parcel] = x
    hh.y[idx_hh_on_parcel] = y


    

    




Fetching parcels
Fetching modify_table
Fetching buildings
Fetching establishments
Fetching modify_table
Fetching modify_table
Fetching households_for_estimation
Fetching modify_table
Fetching households
Fetching modify_table
Fetching zones
Fetching modify_table
Fetching travel_data
Fetching modify_table
Fetching parcel_coords
Fetching modify_table

In [2]:

    

hh.describe()


    

    
Out[2]:






  

    

      

      
household_id
      
building_id
      
parcel_id
      
x
      
y
      
taz05_id
      
dist_trans
    
  
  

    

      
count
      
 1160869.000000
      
 1160869.000000
      
 1160869.000000
      
 1160869.000000
      
 1160869.000000
      
 1160869.000000
      
 1160869.000000
    
    

      
mean
      
  580435.179291
      
  473059.299294
      
    1468.339839
      
 3145353.709304
      
 1699498.501644
      
     245.442482
      
       0.931423
    
    

      
std
      
  335114.414025
      
  294924.760137
      
  314365.345610
      
   45183.335625
      
   62198.538595
      
  170944.540782
      
       2.704356
    
    

      
min
      
       1.000000
      
       1.000000
      
       1.000000
      
 2890124.000000
      
 1472247.000000
      
  101010.000000
      
       0.003409
    
    

      
25%
      
  290218.000000
      
  191967.000000
      
  238158.000000
      
 3116973.000000
      
 1661996.000000
      
  211200.000000
      
       0.097727
    
    

      
50%
      
  580435.000000
      
  466724.000000
      
  563020.000000
      
 3145512.000000
      
 1690691.000000
      
  404450.000000
      
       0.185227
    
    

      
75%
      
  870652.000000
      
  711175.000000
      
  776663.000000
      
 3173554.000000
      
 1735880.000000
      
  506240.000000
      
       0.412121
    
    

      
max
      
 1160870.000000
      
 1015771.000000
      
 1128921.000000
      
 3503103.000000
      
 1885594.000000
      
  803120.000000
      
      51.506818
    
  

In [ ]:

    

e = establishments.reset_index()
for idx in e.index:
    for job in range(e.employees[idx]):
        bids.append(e.building_id[idx])
        eids.append(e.index[idx])
        hbs.append(e.home_based_status[idx])
        sids.append(e.sector_id[idx])
print len(bids)
print len(eids)
print len(hbs)
print len(sids)
jobs = pd.DataFrame({'job_id':range(1,len(bids)+1),'building_id':bids,'establishment_id':eids,'home_based_status':hbs,'sector_id':sids})
jobs['parcel_id'] = bpz.parcel_id[jobs.building_id].values
jobs['urbancenter_id'] = bpz.urbancenter_id[jobs.building_id].values
jobs['x'] = bpz.centroid_x[jobs.building_id].values.astype('int64')
jobs['y'] = bpz.centroid_y[jobs.building_id].values.astype('int64')
big_parcel_ids_with_jobs = np.unique(jobs.parcel_id[np.in1d(jobs.parcel_id,big_parcels)].values)
for parcel_id in big_parcel_ids_with_jobs:
    idx_jobs_on_parcel = np.in1d(jobs.parcel_id,[parcel_id,])
    coords = parcel_coords[parcel_coords.parcel_id==parcel_id]
    idx_coord = np.random.choice(coords.index,size=idx_jobs_on_parcel.sum(),replace=True)
    x = coords.x.loc[idx_coord].values
    y = coords.y.loc[idx_coord].values
    jobs.x[idx_jobs_on_parcel] = x
    jobs.y[idx_jobs_on_parcel] = y


    

In [ ]:

    

jobs.describe()