Creating a cross-tabulated table from data frame in R

Question

I have a data frame like this:

df <- data.frame(Country = rep(c("US","CA"),each=3),
                 Variable = c("Inflation","Unemployment","Interest rate"),
                 Month = rnorm(6), Quarter = rnorm(6)+2, Year=rnorm(6)+3)

and I would like to transform it into something like this:

I have tried the tables package, but did not succeed. Does anyone have any clue about how to achieve this?

EDIT: The output may be either in LaTeX or Text form.

Answer 1

You can't merge data.frame cells in R like you would in excel, but you can combine the Country and period columns. Here's a solution with dplyr + tidyr:

library(dplyr)
library(tidyr)

df %>%
  gather(var, value, Month:Year) %>%
  unite("var", Country, var) %>%
  spread(var, value)

Result:

       Variable   CA_Month CA_Quarter  CA_Year   US_Month US_Quarter  US_Year
1     Inflation  0.2760235   1.758310 4.233976 -0.4321298  3.6232025 5.149919
2 Interest rate -0.5208693   1.227022 3.412022  1.2283928  3.6858872 3.495870
3  Unemployment -1.0489755   1.531800 3.634362  1.6898725  0.9299318 1.665646

To actually create a table (for reporting) that has the Country merged, here is how to do it with a one-liner using the tables package:

library(tables)

tabular(Variable ~  Heading()*Country*Heading()*identity*(Month + Quarter + Year), data=df)

Result:

               CA                    US                   
 Variable      Month   Quarter Year  Month   Quarter Year 
 Inflation      0.5269 2.152   3.854 -0.9456 3.764   1.432
 Interest rate  1.3974 1.820   3.340  0.4520 1.734   3.962
 Unemployment  -0.2303 3.377   3.419 -0.6652 2.486   2.739

tabular uses its unique expression to generate the table formats:

• ~ separates the expression for rows from columns. I am displaying the rows of Variable.

• * means that you are nesting one column into another. In this case, I am nesting the Month:Year columns in Country.

• identity specifies that you are displaying the actual values in each cell.

• Heading replaces the heading of the next item with a string. In this case, I am replacing "Country" and "identity" with blank.

To output as latex, you can wrap the entire expression with the latex function:

latex(tabular(Variable ~  Heading()*Country*Heading()*identity*(Month + Quarter + Year), data=df))

Result:

\begin{tabular}{lcccccc}
\hline
 & \multicolumn{6}{c}{Country} \\ 
 & \multicolumn{3}{c}{CA} & \multicolumn{3}{c}{US} \\ 
Variable  & Month & Quarter & Year & Month & Quarter & \multicolumn{1}{c}{Year} \\ 
\hline
Inflation  & $\phantom{-}0.5269$ & $2.152$ & $3.854$ & $-0.9456$ & $3.764$ & $1.432$ \\
Interest rate  & $\phantom{-}1.3974$ & $1.820$ & $3.340$ & $\phantom{-}0.4520$ & $1.734$ & $3.962$ \\
Unemployment  & $-0.2303$ & $3.377$ & $3.419$ & $-0.6652$ & $2.486$ & $2.739$ \\
\hline 
\end{tabular}

Answer 2

We could try with kable from knitr after reshaping with data.table

library(data.table)
library(knitr)
library(kableExtra)

dt <- dcast(setDT(df),  Variable ~ Country, value.var = c('Month', 'Quarter', 'Year'))
nm1 <- names(dt)
nm2 <- c(" ", unique(sub(".*_", "", nm1)[-1]))

setnames(dt, sub("_.*", "", nm1))
setcolorder(dt, order(ave(seq_along(dt), names(dt), FUN = seq_along)))

kable(dt, 'html') %>%  
   kable_styling('striped') %>% 
    add_header_above(c(' ' = 1, 'CA' = 3, 'US' = 3))

• 'html' table output

<table class="table table-striped" style="margin-left: auto; margin-right: auto;">
<thead>
<tr>
<th style="border-bottom:hidden" colspan="1"></th>
<th style="text-align:center; border-bottom:hidden; padding-bottom:0; padding-left:3px;padding-right:3px;" colspan="3"><div style="border-bottom: 1px solid #ddd; padding-bottom: 5px;">CA</div></th>
<th style="text-align:center; border-bottom:hidden; padding-bottom:0; padding-left:3px;padding-right:3px;" colspan="3"><div style="border-bottom: 1px solid #ddd; padding-bottom: 5px;">US</div></th>
</tr>
<tr>
<th style="text-align:left;"> Variable </th>
   <th style="text-align:right;"> Month </th>
   <th style="text-align:right;"> Quarter </th>
   <th style="text-align:right;"> Year </th>
   <th style="text-align:right;"> Month </th>
   <th style="text-align:right;"> Quarter </th>
   <th style="text-align:right;"> Year </th>
  </tr>
</thead>
<tbody>
<tr>
<td style="text-align:left;"> Inflation </td>
   <td style="text-align:right;"> -0.5836272 </td>
   <td style="text-align:right;"> 2.0023119 </td>
   <td style="text-align:right;"> 2.530939 </td>
   <td style="text-align:right;"> -0.5458808 </td>
   <td style="text-align:right;"> 2.444585 </td>
   <td style="text-align:right;"> 2.237786 </td>
  </tr>
<tr>
<td style="text-align:left;"> Interest rate </td>
   <td style="text-align:right;"> 0.2660220 </td>
   <td style="text-align:right;"> 2.5982691 </td>
   <td style="text-align:right;"> 4.536252 </td>
   <td style="text-align:right;"> 0.4196231 </td>
   <td style="text-align:right;"> 1.151630 </td>
   <td style="text-align:right;"> 3.332244 </td>
  </tr>
<tr>
<td style="text-align:left;"> Unemployment </td>
   <td style="text-align:right;"> 0.8474600 </td>
   <td style="text-align:right;"> 0.6830919 </td>
   <td style="text-align:right;"> 2.665013 </td>
   <td style="text-align:right;"> 0.5365853 </td>
   <td style="text-align:right;"> 1.533505 </td>
   <td style="text-align:right;"> 1.570910 </td>
  </tr>
</tbody>

Note that replacing the 'html' with 'latex' creates the latex output

kable(dt, 'latex') %>%
   kable_styling('striped') %>% 
   add_header_above(c(' ' = 1, 'CA' = 3, 'US' = 3))

-latex output

\begin{table}[H]
\centering
\begin{tabular}{l|r|r|r|r|r|r}
\hline
\multicolumn{1}{c|}{ } & \multicolumn{3}{|c|}{CA} & \multicolumn{3}{|c}{US} \\
\cline{2-4} \cline{5-7}
Variable & Month & Quarter & Year & Month & Quarter & Year\\
\hline
Inflation & -0.5836272 & 2.0023119 & 2.530939 & -0.5458808 & 2.444585 & 2.237786\\
\hline
Interest rate & 0.2660220 & 2.5982691 & 4.536252 & 0.4196231 & 1.151630 & 3.332244\\
\hline
Unemployment & 0.8474600 & 0.6830919 & 2.665013 & 0.5365853 & 1.533505 & 1.570910\\
\hline
\end{tabular}
\end{table}