Code for Historgrams

[Kerrowman]

Hi there,

I'm new to R and R-Studio but have been watching lots of videos on how to get started and have picked up the main gist of how it works.

To practice I have assembled a sample data table in Numbers and exported it as a csv to R and the first image shows the table in R. It lists CoP values (and uncertainties) for 20 test events for two types of test, a regular and a control. I am just practicing showing the first column (CoP) against ID in a histogram.

When I use the code:

ggplot(aes(ID)) + geom_histogram(binwidth = 2, fill = "steelblue", alpha = 0.5) + labs(title = "CoP vs ID",
x = "ID", y = "CoP") + theme_bw()```

it produces a histogram as shown in the second image that does not seem to distinguish the various nuanced values (e.g. 2.12, 2.63, etc) but instead shows CoP values of 1, 2 or 3 against the test ID.

Is this because I have not specified a resolution to the mapping or something similar?

https://ibb.co/vwH7nz6

https://ibb.co/xLsk1sD

[Kerrowman]

I'm not sure I have uploaded the images correctly as I can't get them to show. I will try ImgBB codes

[technocrat]

library(ggplot2)
# default binwdith
mtcars |> ggplot(aes(mpg)) +
  geom_histogram() +
  theme_minimal()
#> `stat_bin()` using `bins = 30`. Pick better value with `binwidth`.

# finer grained
mtcars |> ggplot(aes(mpg)) +
  geom_histogram(binwidth = 0.25) +
  theme_minimal()

^{Created on 2024-01-08 with reprex v2.0.2}

[Kerrowman]

Hi there,

I have tried your code and it makes no difference. In this example, I want the test ID on the X axis and the CoP value on the Y axis. What I get is in the image link

COP1 %>%
  ggplot(aes(ID)) +
  geom_histogram() +
  theme_minimal()

Co-P-plot-2 hosted at ImgBB

It seems I need to specify the Y-axis variable and perhaps some resolution? The histogram blocks are all the same size indicating the frequency of occurrence instead of the CoP value (e.g. 2.14) for each test ID.

[technocrat]

The graph linked isn't a standard histogram which shows the count by interval.

It would help to have a reprex (see the FAQ) to see the data (or similar data) that you are working with.

[Kerrowman]

I showed the COP data table at the start. Here is the link again.

Co-P-Table hosted at ImgBB

I know Histograms are usually categorical on the X-axis and then a count or value on the Y, and I can produce a scatter plot (Co-P-Scatter-plot hosted at ImgBB — ImgBB) but wanted a visual showing Test ID on the X and the COP value as the bar height (Y value). Is that not doable?

Thanks

[technocrat]

You may get more answers if you use something that can be cut and pasted, as described in the FAQ. the dput() function is good to bring data in.

[Kerrowman]

Sounds like a good idea but I can't see how to do this. I have put 'dput' into FAQ and it comes up with how to provide a subset of data that is already loaded into R.

Do you have a link on how to bring in a sample of external data?

[Kerrowman]

Being unable to get 'dput' to work, here is a data frame with the first 10 values of my larger dataset:

Df <- data.frame(ID = c(1, 2, 3, 4, 5, 6, 7, 8, 9, 10), CoP = c(2.12, 2.44, 1.97, 2.63, 2.04, 1.95, 2.52, 2.92, 2.56, 2.17))

All I am trying to do I give an alternative visualization to a scatter graph by showing the ID on the X-Axis and the value for CoP as the height of the histogram.

[nirgrahamuk]

I think you are confusing histogram with a bar chart, a histogram is a very particular subset of the more general bar charts. Do you require binning ? (histogram) , or do you already have binned data (bar chart).
Your ID's to use on the x-axis strongly imply you have a bar chart to generate and therefore geom_histogram is not the approriate tool. if you want the bar to be as high as CoP then you would use geom_col, withaes using y=CoP

[jrkrideau]

Here is a very simple example of how to do it.

dat <- data.frame(xx = 1:10, yy = letters[1:10])

dput(dat)

This gives us

structure(list(xx = 1:10, yy = c("a", "b", "c", "d", "e", "f", 
                                 "g", "h", "i", "j")), class = "data.frame", row.names = c(NA,  -10L))

We can then copy it into R

my_dat  <- structure(list(xx = 1:10, yy = c("a", "b", "c", "d", "e", "f", 
                                            "g", "h", "i", "j")), class = "data.frame", row.names = c(NA,  -10L))

and we have an exact copy of your data.

[Kerrowman]

Clearly I am rusty on the difference: histogram for probabilities and bar charts for values.

This now works:

ggplot(aes(ID, CoP)) + geom_col(fill = "steelblue", alpha = 0.5) + labs(title = "CoP vs ID", x = "ID", y = "CoP") + theme_bw()

Thanks

[Kerrowman]

Quite a bit to get used to there for a beginner but thankfully I don't need that now. Thanks anyway.

[jrkrideau]

Do not worry, if you are used to SAS , SPSS, etc. the sense of complete disorientation fade after a few weeks.

[Kerrowman]

Good to hear that is the case