CODEWITHSUNDEEP
rdata-science
user13398810
user13398810

Reputation:

Unable to plot a data.frame correctly in r

I'm unable to do it by myself, trying it for at least 2 hours...

I have the following data.frame imported from a CSV-file:

> asyl_continents
          2010 2011 2012 2013 2014  2015  2016
Europe    1411 1352 2047 3277 5105  7647  5296
Asia      2832 3229 3903 4418 7123 32152 18002
America     20   17   20   27   35    42    97
Africa     436  757  897 1497 3846  5412  4592
Australia    0    0    0    0    0     0     1
stateless   34   34   35   28   80    88   179
unknown    159  160  140  207  360   408   505

This is the doutput of it:

structure(list(`2010` = c(1411L, 2832L, 20L, 436L, 0L, 34L, 159L
), `2011` = c(1352L, 3229L, 17L, 757L, 0L, 34L, 160L), `2012` = c(2047L, 
3903L, 20L, 897L, 0L, 35L, 140L), `2013` = c(3277L, 4418L, 27L, 
1497L, 0L, 28L, 207L), `2014` = c(5105L, 7123L, 35L, 3846L, 0L, 
80L, 360L), `2015` = c(7647L, 32152L, 42L, 5412L, 0L, 88L, 408L
), `2016` = c(5296L, 18002L, 97L, 4592L, 1L, 179L, 505L)), class = "data.frame", row.names = c("Europe", 
"Asia", "America", "Africa", "Australia", "stateless", "unknown"
))

I want to plot something like this: https://i.sstatic.net/l7s6l.png On the x-achsis there should be the years and for every category (Europe, Asia, ..., unknown) a seperate row.

Wondering what this is about? For a student-project I want to show that crimerates only gone up in absolute numbers with more refugees arriving, but not in relative numbers. Want to conbine this dataset later on with crimerates.

Upvotes: 0

Views: 372

Answers (3)

Martin
Martin

Reputation: 594

The following code does not depend on non-base-R packages.

A) graphics::matplot plots the columns of a matrix, to which the data-frame is coerced toautomatically:

matplot(x = as.numeric(colnames(asyl_continents)), # x values: the years
    y    = t(asyl_continents),                 # y values: one column per plotted line each, therefore transpose years' asylum values to columns 
    # col= c(...)                              # could specify your own colors here, as well as line types, thickness etc.
    type = "l",                                # specify type: draw lines, see ? matplot for details
    xlab = "years", ylab = "asylum")

B) legend() adds a legend

legend(x="topleft", legend = row.names(asyl_continents), 
       col = seq_len(ncol(asyl_continents)),      # use matplot's default color sequencs, unless spefified otherwise in matplot() (see https://stackoverflow.com/questions/27796583/how-to-add-colour-matched-legend-to-a-r-matplot)
       lwd = 1, cex=0.85)

You can change colors, line types etc. as you want, see ? matplot for details.

enter image description here

Upvotes: 1

Peter
Peter

Reputation: 12739

This should set you off in the tidyverse direction to achieve the line graph you aspire to with your data.


library(tibble)
library(dplyr)
library(tidyr)
library(stringr)
library(ggplot2)


# Provide a sample of data in a proper dataframe or equivalent object

df <- tibble(cont = c("Europe", "Asia", "America", "Africa", "Australia", "stateless", "unknown"), 
             `2010` = c(1411L, 2832L, 20L, 436L, 0L, 34L, 159L), 
             `2011` = c(1352L, 3229L, 17L, 757L, 0L, 34L, 160L),
             `2012` = c(2047L, 3903L, 20L, 897L, 0L, 35L, 140L), 
             `2013` = c(3277L, 4418L, 27L, 1497L, 0L, 28L, 207L),
             `2014` = c(5105L,7123L, 35L, 3846L, 0L, 80L, 360L),
             `2015` = c(7647L, 32152L, 42L, 5412L, 0L, 88L, 408L), 
             `2016` = c(5296L, 18002L, 97L, 4592L,1L, 179L, 505L))

# in this case it is helpful to put the data in to longer format 
df1 <- 
  df %>% 
  pivot_longer(cols = matches("\\d{4}$"), names_to = "year", values_to = "asyl_nr")

# plot the data
ggplot(df1, aes(year, asyl_nr, colour = cont, group = cont))+
  geom_line()+
  theme_classic()

And this is what you end up with:

enter image description here

Upvotes: 0

lrnv
lrnv

Reputation: 1106

Welcome to stackoverflow. Please describe a little more what you want as a barplot ? Also, please give us the output of dput(asyl_continents) so we can work with your data frame directly. The following code is 'by memory', i did not test it but i think i'll work.

Something like boxplot(asyl_continents) ?

by relative number, you mean rowwise increasing ? indeed, this is meaningfull : 

df <- asyl_continents
df <- apply(df,1,function(x){1+ (x-x[1])/x[1]})
matplot(t(df))

Upvotes: 0

Related Questions