Re-arranging data for GLM analysis in R using a for-loop

Question

I think my question is fairly simple to answer but I'm learning R so I'd like to know the best way to do it.

I've a dataset looking like this:

> print(agg_df41367)
# A tibble: 72 x 3
# Groups:   hour [24]
    hour predicted      y
 1     0 Feeding   0.121 
 2     0 Foraging  0.632 
 3     0 Standing  0.300 
 4     1 Feeding   0.141 
 5     1 Foraging  0.727 
 6     1 Standing  0.183 
 7     2 Feeding   0.0932
 8     2 Foraging  0.817 
 9     2 Standing  0.133 
10     3 Feeding   0.214 

I would like to run a GLM model, so I'd like my data to look like:

head(agg_df41361_GLM)
hour Foraging Standing Feeding 
0     0.632   0.300    0.121
1     0.727   0.183    0.141
2     0.817   0.133    0.0932
3     etc.    etc.      0.214

Any ideas of what is the most compact way to do this? Ideally, I would like to use a for -loop to compute this transformation for multiple datasets. All my datasets follow a name format agg_df4136* . Any input is appreciated!

Answer 1

Here's a way to reshape the dataset you posted.

library(tidyr)

# example data
dt = read.table(text = "
hour predicted      y
1     0 Feeding   0.121 
2     0 Foraging  0.632 
3     0 Standing  0.300 
4     1 Feeding   0.141 
5     1 Foraging  0.727 
6     1 Standing  0.183 
7     2 Feeding   0.0932
8     2 Foraging  0.817 
9     2 Standing  0.133 
", header=T)

spread(dt, predicted, y)

#   hour Feeding Foraging Standing
# 1    0  0.1210    0.632    0.300
# 2    1  0.1410    0.727    0.183
# 3    2  0.0932    0.817    0.133

If you have multiple datasets it's better to create a list of them and apply the reshaping process to each one of them:

library(tidyverse)

# example of list of dataframes
l = list(dt, dt, dt)

map(l, ~spread(., predicted, y))

# [[1]]
# hour Feeding Foraging Standing
# 1    0  0.1210    0.632    0.300
# 2    1  0.1410    0.727    0.183
# 3    2  0.0932    0.817    0.133
# 
# [[2]]
# hour Feeding Foraging Standing
# 1    0  0.1210    0.632    0.300
# 2    1  0.1410    0.727    0.183
# 3    2  0.0932    0.817    0.133
# 
# [[3]]
# hour Feeding Foraging Standing
# 1    0  0.1210    0.632    0.300
# 2    1  0.1410    0.727    0.183
# 3    2  0.0932    0.817    0.133

Note that here I'm using the same dataset ( dt ) as my 3 list elements, but it will work with different datasets, as long as you have the same column names.

If you want to create a list of all your datasets that start with the name pattern you provided you can do this:

# get objects that start with this name pattern
input_names = ls()[grepl("^agg_df4136", ls())]

# get the data that match those names
list_datasets = map(input_names, get)

So, list_datasets is a list of all dataframes in your environment with a name that starts with "agg_df4136".