In dantonnoriega/sppedr: Imports, cleans, and combines different data sources for Chicago Public Schools and Police Departments

knitr::opts_chunk$set(
  warning = FALSE,
  message = FALSE,
  collapse = TRUE,
  echo = TRUE,
  cache = FALSE,
  comment = "#>",
  fig.path = "README-"
)

Installation

install.packages('devtools') # install devtools if needed
devtools::install_github("dantonnoriega/sppedr") # install the package

Prerequisites

This package assumes you have access to a directory with the following files and structure

spped
└── Data
    ├── cps_2008_2016.dta
    ├── cps_address_2008_2016.dta
    RawData
    ├── CPS_Personnel
    ├── Chicago_Crime
    ├── Crosswalks
    ├── ISBE_Directories
    └── ISBE_School_Data

The directory spped is considered the data "main directory" (or main_dir) by the package. Wherever this is located on your computer, ensure the package knows where it is by running the set_main_dir() command.

library(sppedr)

clear_main_dir() # clear any existing main_dir
try(get_main_dir()) # no main dir
set_main_dir("/Users/danton/Dropbox/ra-work/spped")
# set_main_dir() # can also run without path and it will prompt
get_main_dir()

This will allow you to run the package functions from scratch should you want to generate new data.

Package Datasets

The package, when loaded, also attaches the following datasets. These data sets can be called directly once you have loaded library(sppedr).

library(sppedr)
data(package = 'sppedr') # list of all available data

Dataset Descriptions

cps_2008_2016: school data for years 2008 to 2016. this is a csv version of the dta file cps_2008_2016.dta
cps_school_year: this contains dates about the academic school year. it is a list object with two data frames, cps_school_year$school_year and cps_school_year$days_off.
cps_address: contains lat/lon coordinates and full addresses as understood by the google maps API.
cps_security_personnel
cps_crosswalk: this file helps link the cps_security_personnel data with the all other data sources by merging on the unit_number variable.
cpd_crime_codes: contains crime codes from the Chicago PD (IUCR codes) and the FBI. it is a list of two data frames, cpd_crime_codes$iucr and cpd_crime_codes$fbi.

Example: `cps_address` dataset

# first 10 full school address for academic year 2009 - 2010
cps_address %>% 
  dplyr::filter(school_year == 2010) %>%
  dplyr::select(schoolname, school_year, address_full, lat, lon) %>%
  head(10)

Other Info

In the package directory data-raw, there is a README.txt file that explains where all the internal datasets come from and/or how they are built. Scripts that build most of these data are also in data-raw. Note that this directory is accessible via the github repo, not the package itself.

Functions

The only function that one really needs call is get_cpd_crime(). This is because these data are much larger and could not be included as part of the package. It will import these data from the main_dir set using the set_main_dir() function.

The following functions do not need to be called unless you want to regenerate these data again from scratch.

get_cps_2008_2016()
get_cps_address()
get_cps_security_personnel()

To do so, you need to set force = TRUE. Otherwise, the function will just load the package version. I would recommend NOT to use the force option.

Example

Here is an example where the CPS security personnel data is spread and the merged to the crosswalk data.

personnel_wide <- cps_security_personnel %>%
  dplyr::left_join(., cps_crosswalk, by = 'unit_number') %>%
  dplyr::mutate(job_description = gsub('school ', '', job_description)) %>%
  dplyr::mutate(job_description = gsub('[[:space:]]', '_', job_description)) %>%
  dplyr::filter(!is.na(schoolidr_c_d_t_s)) %>%
  tidyr::spread(job_description, n) %>%
  dplyr::mutate_if(is.integer, dplyr::funs(replace(., is.na(.), 0))) %>% # replace NAs with 0
  dplyr::select(-unit_number, -abbreviated_name) %>%
  dplyr::mutate(school_year = as.integer(school_year))

# list first 10 security_officer counts
personnel_wide %>%
  dplyr::select(school_year, school_name, security_officer) %>%
  head(10)