parcels.py

"""
parcels.py
Spring 2022 PJW

Analyze residential property in Onondaga County by zip code.
"""

import pandas as pd
import matplotlib.pyplot as plt

#
#  Read the parcel data being careful not to clobber the GEOID
#

use_type = {'ZCTA5CE10':str}
raw = pd.read_csv('onondaga-tax-parcels.zip',dtype=use_type)
print( len(raw) )

#
#  Create a couple of variables for convenience
#

raw['zipcode'] = raw['ZCTA5CE10']
raw['value_k'] = raw['TOTAL_AV']/1000

#%%

#
#  Drop records that are missing any of the key variables below
#

key_vars = ['zipcode','value_k','PROP_CLASS']
usable = raw.dropna(subset=key_vars)
print( len(usable) )

#
#  Filter down to residential property
#

is_res = usable['PROP_CLASS'].between(200,299)
res = usable[ is_res ]

#%%

#
#  Compute some overall statistics
#

tot_n = len(res)
mean_k = res['value_k'].mean()
tot_bil = res['value_k'].sum()/1e6

print( 'Residential parcels:', tot_n )
print( 'Mean value, thousand $:', round(mean_k,2) )
print( 'Total value, billion $:', round(tot_bil,2) )

#%%

#
#  Group by zip code, build a dataframe with the medians of 
#  selected variables, and print it out
#

by_zip = res.groupby('zipcode')

attr = ['SQFT_LIV', 'NBR_BEDRM', 'YR_BLT', 'CALC_ACRES']
summary = by_zip[attr].median()
summary = summary.round(2)
summary = summary.sort_values('YR_BLT')

print( summary )

#%%

#
#  Now look at values by zip code in more detail
#

value = by_zip['value_k'].describe()
value = value.round(2)
value = value.sort_values('50%')

print( value.sort_values('50%') )

#%%

# 
#  Make a nice summary dataframe and save it as a csv file
#

for c in value.columns:
    summary[c] = value[c]

summary = summary.sort_index()
summary.to_csv('parcels.csv')

#%%

#
#  Now draw a plot showing three characteristics of housing by 
#  zip code: year built, square feet of living area, and median 
#  value; dots will be scaled by value
#

(fig1, ax1) = plt.subplots()
summary.plot.scatter('YR_BLT','SQFT_LIV',s='50%',ax=ax1)
ax1.set_title('Characteristics of Zip Codes')
fig1.tight_layout()
fig1.savefig('parcels.png',dpi=300)