I'm going to skip the part where you want to transpose, and infer that your purpose for that was solely to help with plotting. The part I'm focusing on here is "show the Month-1, Month-2 etc in x-axis, and the series for each company in the same graph".
This is doable in base graphics, but I highly recommend using ggplot2 (or plotly or similar), due to its ease of dealing with dimensional plots like this. The "grammar of graphics" (which both tend to implement) really prefers data like this be in a "long" format, so part of what I'll do is convert to this format.
First, some data:
set.seed(2)
months <- paste0("Month", 1:30)
companies <- paste0("Comp", 1:5)
m <- matrix(abs(rnorm(length(months)*length(companies), sd=1e3)),
nrow = length(companies))
d <- cbind.data.frame(
Company = companies,
m,
stringsAsFactors = FALSE
)
colnames(d)[-1] <- months
str(d)
# 'data.frame': 5 obs. of 31 variables:
# $ Company: chr "Comp1" "Comp2" "Comp3" "Comp4" ...
# $ Month1 : num 896.9 184.8 1587.8 1130.4 80.3
# $ Month2 : num 132 708 240 1984 139
# $ Month3 : num 418 982 393 1040 1782
# $ Month4 : num 2311.1 878.6 35.8 1012.8 432.3
# (truncated)
Reshaping can be done with multiple libraries, including base R, here are two techniques:
library(data.table)
d2 <- melt(as.data.table(d), id = 1, variable.name = "Month", value.name = "Cost")
d2[,Month := as.integer(gsub("[^0-9]", "", Month)),]
d2
# Company Month Cost
# 1: Comp1 1 896.91455
# 2: Comp2 1 184.84918
# 3: Comp3 1 1587.84533
# 4: Comp4 1 1130.37567
# 5: Comp5 1 80.25176
# ---
# 146: Comp1 30 653.67306
# 147: Comp2 30 657.10598
# 148: Comp3 30 549.90924
# 149: Comp4 30 806.72936
# 150: Comp5 30 997.37972
library(dplyr)
# library(tidyr)
d2 <- tbl_df(d) %>%
tidyr::gather(Month, Cost, -Company) %>%
mutate(Month = as.integer(gsub("[^0-9]", "", Month)))
I also integerized the Month, since it made sense with an ordinal variable. This isn't strictly necessary, the plot would just treat them as discretes.
The plot is anti-climactically simple:
library(ggplot2)
ggplot(d2, aes(Month, Cost, group=Company)) +
geom_line(aes(color = Company))
Bottom line: I don't think you need to worry about transposing your data: doing so has many complications that can just confuse things. Reshaping is a good thing (in my opinion), but with this kind of data is fast enough that if your data is stored in the wide format, you can re-transform it without too much difficulty. (If you are thinking about putting this in a database, however, I'd strongly recommend you re-think "wide", your db schema will be challenging if you keep it.)
