Make scatter (or X, Y) plot by treatment for different time period

Question

I have a data (R dataframe) like this:

Treatment   Diameter(inches).Sep    Diameter(inches).Dec
Aux_Drop    NA  NA
Aux_Spray    3.7    2
DMSO    NA  NA
Water   4.2 2
Aux_Drop    2.6 3
Aux_Spray    3.7    3
DMSO    4   2
Water   5.2 1
Aux_Drop    5.4 2
Aux_Spray    3.4    2
DMSO    4.8 2
Water   4.2 2
Aux_Drop    4.7 2
Aux_Spray    2.7    2
DMSO    3.4 2
Water   4.9 2
.......
.......

I want to make a scatter (or x, y) plot of diameter for each treatment group. I have found lattice library plot more helpful as of now and I have used:

require(lattice)
xyplot(`Diameter(inches).Sep` ~ Treatment , merged.Sep.Dec.Mar, pch= 20)

to generate the plot:

However, I want to add the scatter plot for "Diameter from Dec" next to the "Diameter of Sep" for each treatments with different color. I am not able to find a workable example that I can use for my purpose so far.

Method with lattice, ggplot2 or base plot or any other would be really helpful.

Thanks,

Your example data seems to have 4 columns but only 3 headers? — neilfws, Sep 12 '18 at 23:03
Oops bummer. I am going to remove that. Thanks. But, it really doesn't affect the intent of the question. — everestial007, Sep 12 '18 at 23:05
No, but it does make it difficult to answer when we try to recreate your data :) — neilfws, Sep 12 '18 at 23:06
Yeah, totally. That was mistake. Just fixed it. Thanks for pointing out. — everestial007, Sep 12 '18 at 23:07
Possible duplicate of [Side-by-side plots with ggplot2](https://stackoverflow.com/questions/1249548/side-by-side-plots-with-ggplot2) — divibisan, Sep 13 '18 at 00:37

score 1 · Accepted Answer · answered Sep 12 '18 at 23:14

1

Something like this?

library(tidyverse)
df %>%
    gather(Month, Diameter, -Treatment) %>%
    ggplot(aes(Treatment, Diameter)) +
    geom_point(aes(colour = Month), position = position_dodge(width = 0.9))

You can adjust the amount of separation between the different coloured points by changing width inside position_dodge.

Sample data

df <- read.table(text =
    "Treatment   Diameter(inches).Sep    Diameter(inches).Dec
Aux_Drop    NA  NA
Aux_Spray    3.7    2
DMSO    NA  NA
Water   4.2 2
Aux_Drop    2.6 3
Aux_Spray    3.7    3
DMSO    4   2
Water   5.2 1
Aux_Drop    5.4 2
Aux_Spray    3.4    2
DMSO    4.8 2
Water   4.2 2
Aux_Drop    4.7 2
Aux_Spray    2.7    2
DMSO    3.4 2
Water   4.9 2", header = T)

answered Sep 12 '18 at 23:14

Maurits Evers

49,617
4
47
68

This is totally good. Btw, how did you create new "data or variable" i.e "Month", "Diameter" ? – everestial007 Sep 12 '18 at 23:23
@everestial007 In `gather(Month, Diameter, -Treatment)`, we reshape data across all columns except `Treatment` from wide to long, and in doing so store column names in a new column `Month` and its values in column `Diameter`. – Maurits Evers Sep 12 '18 at 23:26
could you add the code to reshape the data. I tried using other examples but I can't seem to work it out. – everestial007 Sep 13 '18 at 00:00
1

@everestial007 `gather(Month, Diameter, -Treatment)` already reshapes the data from wide to long. It's just the one line. – Maurits Evers Sep 13 '18 at 00:03
Alrite, I got it. I really need to look into how this reshape works. The problem that has happened is that I have several other columns of data for other traits, and this seems to plot everything by treatment. For now, I will reduce the data only by months. Thanks – everestial007 Sep 13 '18 at 00:12

neilfws · Answer 2 · 2018-09-12T23:20:13.723

1

Here's a tidyverse solution. It uses tidyr::gather to put the two diameter types into one column. You can then facet on the values in that column. I hide the colour legend, since the categories are apparent from the axis labels.

Assuming the data frame is named mydata.

library(tidyverse)
mydata %>% 
  gather(Result, Value, -Treatment) %>% 
    ggplot(aes(Result, Value)) + 
    geom_jitter(aes(color = Result), 
                width = 0.1) + 
    facet_wrap(~Treatment) +
    guides(color = FALSE)

edited Sep 12 '18 at 23:20

answered Sep 12 '18 at 23:14

neilfws

32,751
5
50
63

@neifws : This not what I want. I want to have data for say (Aux_Drop) next to each other in the same plot – everestial007 Sep 12 '18 at 23:17
Not, exactly what I intended, but your answer is also helpful for another way of representing the data :) – everestial007 Sep 12 '18 at 23:23

Make scatter (or X, Y) plot by treatment for different time period

2 Answers2

Sample data