CODEWITHSUNDEEP
rinterpolation
rmccoy
rmccoy

Reputation: 13

Interpolate based on multiple conditions in r

Beginner r user here. I have a dataset of yearly employment numbers for different industry classifications and different subregions. For some observations, the number of employees is null. I would like to fill these values through linear interpolation (using na.approx or some other method). However, I only want to interpolate within the same industry classification and subregion.

For example, I have this:

subregion <- c("East Bay", "East Bay", "East Bay", "East Bay", "East Bay", "South Bay")
industry <-c("A","A","A","A","A","B" )
year <- c(2013, 2014, 2015, 2016, 2017, 2002)
emp <- c(50, NA, NA, 80,NA, 300)

data <- data.frame(cbind(subregion,industry,year, emp))

  subregion industry year  emp
1  East Bay        A 2013   50
2  East Bay        A 2014 <NA>
3  East Bay        A 2015 <NA>
4  East Bay        A 2016   80
5  East Bay        A 2017 <NA>
6 South Bay        B 2002  300

I need to generate this table, skipping interpolating the fifth observation because subregion and industry do not match the previous observation. 

  subregion industry year  emp
1  East Bay        A 2013   50
2  East Bay        A 2014   60
3  East Bay        A 2015   70
4  East Bay        A 2016   80
5  East Bay        A 2017 <NA>
6 South Bay        B 2002  300

Articles like this have been helpful, but I cannot figure out how to adapt the solution to match the requirement that two columns be the same for interpolation to occur, instead of one. Any help would be appreciated.

Upvotes: 1

Views: 373

Answers (1)

akrun
akrun

Reputation: 887711

We could do a group by na.approx (from zoo)

library(tidyverse)
data %>% 
     group_by(subregion, industry) %>%
     mutate(emp = zoo::na.approx(emp, na.rm = FALSE))
# A tibble: 6 x 4
# Groups:   subregion, industry [2]
#  subregion industry  year   emp
#  <fct>     <fct>    <dbl> <dbl>
#1 East Bay  A         2013    50
#2 East Bay  A         2014    60
#3 East Bay  A         2015    70
#4 East Bay  A         2016    80
#5 East Bay  A         2017    NA
#6 South Bay B         2002   300

data

data <- data.frame(subregion,industry,year, emp)

Upvotes: 1

Related Questions