Data Visualization for Genome Biology Lab 1

Lab 1: Examining Your Data – A Picture is Worth a Thousand Words

[Software needed: web access]
It is often useful to plot your data in order to be able to identify outliers, to explore differences
between datasets, and to make publication-quality figures. In this lab we’ll explore four datasets.
Table 1: Tools covered in this lab (see end of lab for references)
Tool Notes
Excel/Google Sheets Excel or Google Sheets can provide an easy way of visualizing
your data. We’d quickly run into limitations if we want to make
scientific graphs, but for a quick view these tools are handy.
We’ll introduce some useful keyboard shortcuts to help with
navigation too.
RAWGraphs An “open-source data visualization framework”
R A go-to statistical programming language for biologists with
many packages for working with and plotting data.

Download the Lab1_scatterplot_set_1.txt and three similarly named datasets to your computer
from the course website (typically, right click > Save link as… works for Windows computers; for Mac, hold
down the Control key while clicking to see the local save option. Note that sometimes downloaded files will have an
extra extension added to them and you’ll need to remove this. Search for how to see filename extensions on the web
for your operating system, and then rename the files so that they are called exactly Lab1_scatterplot_set_1.txt etc.)

Excel/Google Sheets
1. Go to Google Sheets at sheets.google.com or start up Excel. To view your data file in
Excel, do File > Open and then identify the Lab1_scatterplot_set_1.txt file you saved on
your local computer. Excel will ask you how you wish to open the file: select Delimited,
then choose “tab” as the delimiter. In Google Sheets, use the File > Import option and use
the defaults. See Figure 1 for a depiction of the first dataset in Google Sheets.

2. Let’s compute the average for the x_values and y_values columns. Find an empty cell and
enter “=average(” and then highlight the cells for which you want to compute the
average. There are a number of ways of doing this. Either type “B2:B143” or use your
mouse to highlight the cells. Alternatively, click the first cell and then hold the Shift key
and highlight the rest of the cells by using the Page Down or down cursor keys. See
Figure 2 on the next page for a list of useful keyboard shortcuts. Enter a closing “)” and
press Enter to see the average value for the x_values column. Do the same for y_values,
and then do the same for the three other datasets.

What are the x_values and y_values averages for each of the 4 datasets? Lab Quiz Question
What are the standard deviations for them? Hint: use the STDEV function! 1
*Answer lab quiz
questions while doing lab!

2023 N.J. Provart, J. Waese, D.S. Guttman, and C. Mok 1

Data Visualization for Genome Biology Lab 1

Figure 1. Examining the first dataset in Google Sheets (Excel view is similar). Compute the average for the
x_values using the =average( ) function, specifying the cells for which to compute the average (here
B2:B143).

Shift plus Page Down select from a cell to page end

Shift then  … add adjacent column to selection

Shift then  … add adjacent row to selection

Shift plus End then  … select from a cell to
[Excel only] end of row
Shift plus End then  … select from a cell to end
[Excel only] of column
Shift plus Space select the whole row
Ctrl plus Space … select the whole column
Ctrl then - [Excel] delete selected
Ctrl plus Alt then - [Sheets] row/column/cell
Figure 2. Useful keyboard shortcuts for navigating in Excel or Google Sheets.
Figure 2. Useful keyboard shortcuts for navigating in Excel or Google Sheets.

2023 N.J. Provart, J. Waese, D.S. Guttman, and C. Mok 2

Data Visualization for Genome Biology Lab 1

3. Let’s plot our data. Move to cell B2 with you cursor keys. Use Ctrl plus Space to
highlight the second column. Use Ctrl plus  to highlight the third column. Then go
Insert > Chart (there are keyboard shortcuts for this, too ͦ). In Google Sheets, change
the Chart type to Scatter chart . In Excel use the Scatter plot icon .

Figure 3. Scatter plot of the first dataset in Google Sheets. The Customize tab has been used to adjust the x
axis minimum and maximum values so that they range from 0 to 100. The same adjustment can be carried
out by clicking on the x axis values in the chart in Excel.

What does each of the scatterplots look like? You can generate these in
Google Sheets or in Excel, or do this in the R section below.

As you can see, it is quite useful to visualize the data using simple charting functions in Excel or
Google Sheets. Other features of these tools include conditional formatting of cells to be able to
provide heatmap-like background colours to cells to be able to easily identify outliers. However,
sometimes we’d like more options for charting. We’ll explore a couple of other tools next.
RAWGraphs
RAWGraphs describes itself as “an open-source data visualization framework built with the goal
of making the visual representation of complex data easy for everyone… providing a missing
link between spreadsheet applications (e.g. Microsoft Excel) and vector graphics editors”.

2023 N.J. Provart, J. Waese, D.S. Guttman, and C. Mok 3

Data Visualization for Genome Biology Lab 1

1. Go to rawgraphs.io and scroll down to see examples of visualizations that are possible in
the “Gallery” section.

2. Click “Use it now”, select “Upload your data”, and drag Lab1_scatterplot_set_1.txt into
the upload section, or select the dataset by browsing to it.

3. In the Choose a chart section (scroll down the page), select “Bubble chart”. This is just
like a scatter plot but you can also map data to the colour and size of dots.

4. In the Mapping section (again, scroll down), drag the x_values and y_values dimensions
onto the “X Axis” and “Y Axis” chart slots, as shown in Figure 4.

Figure 4. Assigning our x_values and y_values dimensions to chart variables. Simply drag the appropriate
dimension to the desired chart variable slot.

5. In the Customize section, explore some of the options, including “Set X origin to 0” and
“Show stroke”, “Max diameter”, and change the colours, as shown in Figure 5.

Figure 5. Changing chart options in RAWGraphs.

6. In the first data set, the points are not exactly at the same positions, so they don’t overlap,
but it’s difficult to see which regions are the densest. Let’s try another chart that is useful
for visualizing distributions and density.

2023 N.J. Provart, J. Waese, D.S. Guttman, and C. Mok 4

Data Visualization for Genome Biology Lab 1

7. Scroll back up to the Choose a chart section and select “Contour plot”. Drag the x and y
dimensions onto the chart slots. View the graph by scrolling down to the plot section.
Check out the customization options in the Customize section, as in Figure 6.

Figure 6. Generating a contour plot in RAWGraphs.

Where is the greatest number of x-y points located? Are you able to figure Lab Quiz
out how to adjust the x axis so that the maximum value is 100? Question 2

8. For the final step, save your chart as an SVG image that you can later use in a
presentation or edit with an SVG image editing program such as Figma, Adobe
Illustrator, or the open-source program Inkscape, from inkscape.org. Use the Export
section to do so.

R
R is a statistical programming environment and possesses many methods for creating graphs and
charts. Although R is complex and can be intimidating, this exercise will serve as a superficial
introduction to how we can recapitulate what we’ve done in a spreadsheet program and with
RAWGraphs with a few keystrokes. Another more detailed tutorial is referenced at the end of this
lab.
You can use your own installation of R as per the Where to get it section, or use Coursera’s built-
in R Notebooks. R is similar to Matlab or SAS. It is dynamically typed so if you have experience
with either MATLAB, SAS, or programming in general, then you should feel pretty comfortable
fairly quickly in R. The most straightforward way to do this lab is to use an R Notebook running
on Coursera’s Jupyter Hub. An R Notebook instance permits you to use a nice web interface to
interact with R. Click on to start up your own Jupyter Hub instance in the section of the

2023 N.J. Provart, J. Waese, D.S. Guttman, and C. Mok 5

Data Visualization for Genome Biology Lab 1

Coursera platform where you found this lab. Your work may not be saved if you close your
browser but you can save your workspace at any time by typing save.image("Lab1.RData") at
the prompt. You can download this file and upload it to continue at a later point.
To instantiate an R Notebook in Jupyter Hub, use the New  > Notebook: R to fire one up, as
shown in Figure 7. Or use the two pre-loaded R Notebooks. The “Lab 1 R (code).ipynb” R
Notebook has all the code shown on the following pages. The “Lab 1 R (sandbox)” is a blank R
Notebook where you can copy or type the code shown on the following pages. We’ve also
provided the datasets (you can upload these on your own instance, see below). Another option is
to use an RStudio interface.

Figure 7. Creating an R Notebook in Jupyter Hub…or use the two pre-loaded .ipynb R Notebooks by clicking on
them.

You’ll see an R Notebook where you can enter commands. Click the character beside a cell to
run the commands in that cell. You can add new cells by clicking the .
. icon. See Figure 8 for a
depiction of a new R Notebook, renamed to “Lab 1” (click on “Untitled” to rename it), with a
command entered into a code cell.
If you don’t have any programming experience, then here are three things about R:
a) Syntax
Like any programming language, the correct use of syntax (also known as language grammar) is
important. This means that every character, variable name, method, parentheses, comma, and
sometimes even space(s) can have importance to the code. Often our toughest enemy can be typos,
as these can create a wide variety of error messages. So, keep a careful eye on how you reproduce

2023 N.J. Provart, J. Waese, D.S. Guttman, and C. Mok 6

Data Visualization for Genome Biology Lab 1

the following steps (you can actually just copy-paste the code if you wish)! If you want to learn
more about R as a data science language, you can peruse the "R for Data Science" e-book by
Hadley Wickham found here: https://r4ds.had.co.nz/introduction.html.
b) Variable assignment
Variable assignment is analogous to the algebraic assignment of variables, e.g. x = 3. However,
the assignment of complex data structures to variables is possible and variable names can be of
any length.

Figure 8. Typing a command into a code cell.

Try entering the following commands into the code cell, as denoted above. Or use the R Console
in the left pane of RStudio (don’t forget to hit Enter after each bolded command if using
RStudio):
mydata = c(3,2,8,1)

mydata

3 2 8 1 #outputs from running a cell with (or by hitting Enter in RStudio)

comments will be in regular weight font

Here we created a variable ‘mydata’ and assigned a vector of 4 numbers to it. A vector in
mathematical terms is simply a series of numbers. As you may have gathered, the typing of the
variable name and hitting enter returns the variables contents. But what about the c( )?
c) Methods
Methods are predefined programs that take data and parameters (placed within parentheses),
perform some calculation or operation on those data (within the context of the provided
parameters) and return a result. This result is usually assigned to a variable but doesn’t have to be.

2023 N.J. Provart, J. Waese, D.S. Guttman, and C. Mok 7

Data Visualization for Genome Biology Lab 1

Here c( ) is the combine method. It combines comma delimited numbers into a vector of numbers
and returns the vector, which we assigned to the variable ‘mydata’. Type ‘?c’ within R to get a
detailed description of the c( ) method.
With this background information in mind, we’re all set to create some graphs in R!

1. First, we’ll load some packages that will help us with our goal of creating a scatterplot. See
supplemental material if you want more information on these...tidyverse is a widely
used package in data science. Not all of these packages will be used today.
# Packages to help tidy our data
library (tidyverse)
# Packages for the graphical analysis section
library (repr)
library (viridis)
# packages used for working with/formatting dates in R
library (lubridate)
library (zoo)

You will receive several messages back about these packages having been attached
successfully, with a few warnings…you can ignore these.

2. Now let’s check out what our working directory is. This is where data will be loaded
from. Create a new code cell by clicking on the .
. icon. Then type or copy the following.
getwd ()

’/home/jovyan/work’#if using Coursera’s Jupyter Hub

To get the above output you can again click the beside the code cell but a handy
shortcut to execute the code cell and to add a new code cell below that is to hold the Alt
key (Option key on a Mac) then hit the Enter key.

3. We’ll have to upload some data from our computer if you’re using your own instance –
if you haven’t done so already, download the Lab1_scatterplot_set_1.txt and three other
datasets from the course website to your computer first, and then click the Upload
button in the top right of your Jupyter Hub to upload the files (hold “Ctrl” to select
multiple files). Click the blue Upload buttons beside each one to finish the process of
uploading.

If you’re using the Coursera R Notebook instance, you will have seen that we have
already provided these files, as shown in Figure 7.

2023 N.J. Provart, J. Waese, D.S. Guttman, and C. Mok 8

Data Visualization for Genome Biology Lab 1

Figure 9. Uploading data to our R Notebook…only required if you’re using your own R Notebook
instance.

4. OK, now we can create a plot! First, load your data.

# Initialize a plot with our data

plot_1.df <- read_tsv("Lab1_scatterplot_set_1.txt")

Rows: 142 Columns: 3

Column specification

Delimiter: "\t"
chr (1): dataset
dbl (2): x_values, y_values
...

5. We can examine the data frame structure using the str command.
# Take a quick look at the structure of the data
str(plot_1.df)

$ dataset: chr [1:142] "set_1" "set_1" "set_1" "set_1" …

$ x : num [1:142] 50.5 50.3 50.2 50.3 50.5 …
$ y : num [1:142] 93.2 97.6 99.7 90 90 …
- attr(*, "spec")=
.. cols(
.. dataset = col_character(),
.. x_values = col_double(),
.. y_values = col_double()
.. )

2023 N.J. Provart, J. Waese, D.S. Guttman, and C. Mok 9

Data Visualization for Genome Biology Lab 1

6. We will use a powerful R package called ggplot to generate a scatter plot based on the
data we put into our data frame.

# Instantiate ggplot object

plot_1.plot <- ggplot(plot_1.df)

There are many different kinds of graphs available. Here’s a selection:

geom_point() for scatter plots

geom_line() for line graphs
geom_boxplot() for boxplots
geom_violin() for violin plots
geom_bar() for bar graphs
geom_histogram() for histograms

7. With the commands shown below, we’ll specify some parameters for our plot, and then
plot it, as shown in Figure 10. The numbered arrows show where you can add some
additional parameters in Step 8.

# Update the aesthetics with axis and colour information, then

add a scatter graph!
plot_1.plot <- plot_1.plot +
aes(x = x_values, y = y_values, colour = dataset) +
geom_point() +
 theme(text = element_text(size = 20)) + # set text size
guides(colour = guide_legend(title="Set 1")) + # Legend title
xlab("x value") + # Set the x-axis label
ylab("y value") + # Set the y-axis label
xlim(0,100) + # specify the minimum and maximum x axis
values
ylim(0,100) # ditto for the y axis

#display our plot
plot_1.plot

2023 N.J. Provart, J. Waese, D.S. Guttman, and C. Mok 10

Data Visualization for Genome Biology Lab 1

Figure 10. Scatter plot of the first dataset with ggplot, in R Notebook.

There are lots of ways to change the aesthetics of the graph. Check out this useful
RStudio cheat sheet: https://raw.githubusercontent.com/rstudio/cheatsheets/main/data-
visualization.pdf. Try choosing a different theme, shapes, or colours, as provided on page
2 of the cheat sheet…see some variants below.
8. View the resulting images for all four graphs. Simply repeat the code in Steps 4 – 7 but
substitute the different file names. It would also be good to create new plot names, so
plot_1.df <- read_tsv("Lab1_scatterplot_set_1.txt") would become
plot_2.df <- read_tsv("Lab1_scatterplot_set_2.txt"). Make sure you
change reference from plot_1 to plot_2 in the new code, so plot_2.plot <-
ggplot(plot_2.df) etc. Last, update the label by changing title="Set 1".

Try adding the following at the indicated locations in Step 7 for plots 3 and 4,
respectively. Rerun Steps 4-6 if you’re just updating the code, before rerunning Step 7.
 theme_bw() + # check out themes at https://ggplot2-book.org/polishing.html

# use a different palette, see https://ggplot2-book.org/scale-colour.html

 plot_3.plot <- plot_3.plot + scale_colour_brewer(palette =
"Pastel2") #colour_brewer palettes are colourblind-safe!
or
 plot_4.plot <- plot_4.plot +
scale_colour_manual(values= c("blue1", "blue2", "blue3", "blue4"))
# we only use the first colour ("blue1") here, as there is only one dataset

If you haven’t done so in the first part of the lab, comment on the
four plots… Lab Quiz
Question 3

2023 N.J. Provart, J. Waese, D.S. Guttman, and C. Mok 11

Data Visualization for Genome Biology Lab 1

9. Last, let’s save our plot. You will be able to find the image in your root directory on
Jupyter Hub.
# Save the plot we've generated to the root directory of the
files.
ggsave(plot = plot_1.plot, filename = "Set_1.png", scale=2,
device = "png", units = c("cm"))

Saving 33.9 x 33.9 cm image

End of Lab!
Lab 1 Objectives
By the end of Lab 1 (comprising the lab including its boxes, and the lectures), you should:
 understand why context is important for data visualization;
 be familiar with the terminology (grammar) used for describing parts of charts;
 know common chart types, and those that are used in biology;
 be able to create simple scatter plots in Excel or Google Sheets;
 be able to appreciate the power (and limitations) of RAWGraphs for creating complex
data visualizations;
 be able to use R and ggplot to graph some data.

Do not hesitate to check with the Forums for this course on Coursera if you do not understand
any of the above after reading the relevant material.

Where to get it: You can run your R visualizations in an R Notebook available within
Coursera’s Jupyter Hub, as described in this lab. If you want to run R on your own computer,
note that this lab was developed with R version 3.6.3, and with a version of dplyr (1.1.0) more
recent than that available in Coursera’s base instance of Tidyverse. We’ve also installed these
packages: RColorBrewer, umap, UpSetR, and viridis. You can download older versions of R
here: https://cran.r-project.org/bin/windows/base/old/. Then start R and use the following
command to add the packages we’re using: install.packages(c('tidyverse',
'RColorBrewer', 'umap', 'UpSetR', 'viridis', 'dplyr', ).

Additional Resources
The four datasets that we used for this lab are a subset of 13 datasets generated by Matejka and
Fitzmaurice in 2017 (see reference below). They all have the same statistical properties and are a follow-
up to Francis Anscombe’s 1973 “quartet” of four datasets that result in very different plots but almost
identical statistical properties.

2023 N.J. Provart, J. Waese, D.S. Guttman, and C. Mok 12

Data Visualization for Genome Biology Lab 1

Justin Matejka and George Fitzmaurice. 2017. Same Stats, Different Graphs: Generating Datasets with
Varied Appearance and Identical Statistics through Simulated Annealing. Proceedings of the 2017 CHI
Conference on Human Factors in Computing Systems. Association for Computing Machinery, New York,
NY, USA, 1290-1294. DOI: https://doi.org/10.1145/3025453.3025912.
Francis Anscombe (1973) Graphs in Statistical Analysis. The American Statistician, 27:1, 17-21, DOI:
http://dx.doi.org/10.1080/00031305.1973.10478966.
RStudio’s ggplot cheat sheet: https://raw.githubusercontent.com/rstudio/cheatsheets/main/data-
visualization.pdf
"R for Data Science" e-book by Hadley Wickham at https://r4ds.had.co.nz/introduction.html

Tips for creating better figures for biology: Rougier NP, Droettboom M, Bourne PE (2014) Ten Simple
Rules for Better Figures. PLoS Comp Biol 10(9): e1003833. https://doi.org/10.1371/journal.pcbi.1003833
Data visualization style guide mentioned on Slide 21 of the Data Visualization Literacy lecture, compiled
by Jamie Waese: Google Docs link

2023 N.J. Provart, J. Waese, D.S. Guttman, and C. Mok 13