census_data_collection

A package improving the data collection process from United States Census Bureaus's API).

The package is built upon census 0.8.19.

Additional features of this package:

• collect yearly (e.g., 2021) or multi-year data (e.g., 2016-2021) from Census Bureaus' API (particularly, American Community Survey) and transform them into structured long-format dataframe automatically
• simplified parameter input

Install

pip install census_data_collection

Usage and Example

First, get yourself a Census API key.

api_key = your_API_key

Then, decide the following parameters:

• estimates method (1-year, 3-year and 5-year estimates, for the differences in estimates, please refer to this document); for example:

est = 1 #1-year estimates
est = 5 #5-year estimate

• variable dictionary: keys are the variable name in Census databases, values are the p re of the variable name you want to rename (the variable name in the databases are meaningless); for example:

var_dict = {'B19013_001E':'income','B01003_001E':'total_population'}

find out all variables using https://api.census.gov/data/[year]/acs/[estimate]/variables.html (e.g., https://api.census.gov/data/2021/acs/acs5/variables.html)

• geographic setting: the geographic area of which data you want to collect (please refer to this file for geographic areas covered in Census; for example:

geo_setting = 'state: *' #collect data from all states
geo_setting = 'county: *' #collect data from all counties

#to get fip codes of each state:
import us
us.states.[state].fips

Now, you are ready to collect initatiate an object for collecting data

from census_data_collection import ACS_data

ACS_data(1, var_dict, geo_setting, api_key)

Next, you may choose to collect single-year data or multi-year data

example = ACS_data(1, var_dict, geo_setting, api_key)

example.collect_data_yearly(2021) #collect 2021 data

import numpy as np
year_range = np.arange(2010, 2016)
example.collect_data_multi_years(np.arange)#collect data from 2010 to 2015

The functions above will return a long-format dataframe (with geo-location and year information):

• the example output for collecting single-year data

• the example output for collecting multi-year data