This repo contains R and Python code that reads the output of DYNASIM's Family and Earnings History (FEH) module.

Python

The Python reader is implemented in the module feh_io. To install it from Github, execute the following command:

pip install feh_io@git+https://github.com/UI-Research/RreadFEH/

The function read_feh_data_file() is all that is needed for reading a DYNASIM file. For example:

import feh_io

header_file = 'C:/Users/dcosic/Documents/Dynasim/Dynasim-core/run/dynasipp_header_even.dat'
person_file = 'C:/Users/dcosic/Documents/Dynasim/Dynasim-core/run/dynasipp_person_even.dat'
family_file = 'C:/Users/dcosic/Documents/Dynasim/Dynasim-core/run/dynasipp_family_even.dat'

# Read 10 records from each person and family file
perdata = feh_io.read_feh_data_file(header_file, person_file, file_type='person', count=10)
famdata = feh_io.read_feh_data_file(header_file, family_file, file_type='family', count=10)

R

For the R reader, its main piece is the code in read_codebook.R that reads DYNASIM's codebook and constructs column names for the data frame.

DYNASIM codebook specifies the logical layout of the person and family records. A record generally includes so called static variables, which contain a single value, and micro time-series, which contain values for a range of years. For each variable on a record, the codebook specifies its name, starting position, and end position. For micro time-series, the codebook also specifies the year range.

Important

All variables are 32-bit integers. The values of starting and ending position of a variable are not used.

A sample codebook is in codebook_2087ds.sipp2006. To obtain column names, execute the following code:

library(tidyverse)
source("read_codebook.R")
rec_struct <- get_col_names(paste0(here::here(), '/codebook_2087ds.sipp2006'))

The read_feh.R file contains code that reads a DYNASIM output file, but this repo does not contain any sample output files.

Data

Sex

Sex is encoded in variable SEX.

• Men: SEX==1
• Women: SEX==2

Race and Ethnicity

Race and ethnicity are encoded in variable ETHNCTY.

• Asian: 4000 <= ETHNCTY <= 4999
• Black: 2000 <= ETHNCTY <= 2999
• Hispanic:
  • 1014 <= ETHNCTY <= 1020 or
  • 2014 <= ETHNCTY <= 2020 or
  • 3014 <= ETHNCTY <= 3020 or
  • 4014 <= ETHNCTY <= 4020
• White: 1000 <= ETHNCTY <= 1999
• Other: 3000 <= ETHNCTY <= 3999

Education

Educational attainment is encoded in the vector GRADECAT as the number of completed grades. It can take values from 0 to 18.