In trottieralex/DHScalendR: Use Demographic Health Survey Calendar Variables

knitr::opts_chunk$set(
  collapse = TRUE,
  comment = "#>",
  fig.path = "man/figures/README-",
  out.width = "100%"
)

DHScalendR

The goal of DHScalendR is to make the use of the DHS calendar easier to use in R. This package is currently being developped, but it already contains functions to:

Deal with CMC date format (back and forth)
Split Calendar into months columns
Recoding of the calendar
Subset using dates (CMC)
Check for combination of events (e.g. postpartum FP use = Birth & any contraception)

Next additions will include ways to:

Locate events
NB. Also need better documentation and further tests.
NB. The type of input will also evolve by version 0.3.0

Installation

To install the current version of this package:

devtools::install_github("trottieralex/DHScalendR")

Data

This package contains a small exemple dataset extracted from a DHS model dataset, and also a codebook to recode the calendar. The later is primarily to use with the calendar_recode function, but is available for consultation or use.

# Exemple dataset
data <- dhs_extract
head(data)
?dhs_extract

# Calendar codebook
head(dhs_calendar_codebook)

Transform CMC date format

The DHS widely use CMC format date (number of month counting from January 1900). A format often convenient for analysis, but hardly interpretable. These function make the transfer CMC from/to Date and years as decimal straigthforward.

library(DHScalendR)

# The Date of the Teharan declaration:
cmc_to_Date(761)
cmc_from_Date(as.Date("1963-05-13")) #Show warning the 1st time

# or for April 19, 1989
cmc_to_Year(1072)
cmc_from_Year(1989.296)

Split and recode DHS calendar

The split function divides the DHS calendar from a data frame into month columns that are added to the said data frame under the name: "CMCx" + [CMC date of the month]. Once split, the recode function coud be use to obtain more meaning full labels or classification.

dataset <- dhs_extract[95:100, c("V008", "V025", "VCAL.1")]

#Dividing months into columns
dataset <- calendar_split(dataset, "V008", "VCAL.1")
(dataset <- dataset[ ,1:10])

#Recoding of calendar months
calendar_recode(dataset$CMCx1326)
calendar_recode(c("1","0","P","P","P","T","A","A","A","A","3","3","3"), 
                recoding = "modern_DHS6")

Subset by dates

A simple way to subset the DHS calendar for a predefined period of time.

dataset <- DHScalendR::dhs_extract[5:9, ]

# Request first half of 2014
calendar_subset(calendar = dataset$VCAL.1, 
                interview = dataset$V008, 
                starting = 1381, 
                ending = cmc_from_Date(as.Date("2015-06-30")))

Check for combination of events

Look for any event2 that hapenned X months or less after any event1.

example <- c("333000TPPP", "000000BPPP", "0000LLLBPPP")
fp <- dhs_calendar_codebook$Recode[dhs_calendar_codebook$modern_DHS6 == "Modern"]

# Check if use any modern FP in the 6 months after end pregnancy
calendar_combo(calendar = example, 
               event1 = c("B", "T"), 
               n_months = 6,
               event2 = fp)