This document provides a brief introduction to the Introductory Biostatistics Labs in R (iblir) package.
The package is not yet on CRAN, so you need to install it from GitHub as follows.
install.packages('remotes')
library('remotes')
install_github('andreashandel/iblir')After the package is installed, you can load it
library('iblir')To start the tutorials/labs, you type
run_tutorial('LabN', package = 'iblir')where N is the lab number, e.g. Lab1,
Lab2.
A startup message also tells you this.
The lab should open in your browser, and you can go through it.
Once finished with a lab, close the tab in your browser. Then go back
to RStudio and stop the lab by pressing the
Stop sign in the upper right corner of the
Background Jobs tabs. You can then go back to the
Console tab and use run_tutorial() again to
start another lab.
If you use RStudio, there is an alternative to calling
run_tutorial(). Once iblir is loaded, you
should see its tutorials listed in the Tutorial tab in the
upper right window.
You can click the Start Tutorial button to start a lab.
It will open in the built-in browser. You’ll likely want to open the lab
in a full browser window by clicking the Show in new window
icon in the top left corner of the tutorial. Note that this method uses
RStudio’s built-in browser, which is not always quite as good as a real
web browser. So if things don’t fully work, switch to the method of
using run_tutorial() and make sure it opens in your real
browser.
LAB 1: Introduction to the online lab format and Introduction to R and “Data”.
LAB 2: Focuses on interpreting and making histograms in R. 1. Interpreting Histograms 2. Which variable types to use with histograms 3. Making histograms 4. Impact of bin selection and 4 methods to help get the right size
LAB 3: Focuses on interpreting and making box plots in R 1. Interpreting Box plots 2. Making Box plots 3. Comparing box plots
LAB 4: Introduces students to the steps of a basic data analysis 1. Data Cleaning 2. Variable types in R 3. Making Tables 4. Making Figures with numeric summaries
LAB 5: Life Tables 1. Life expectancy 2. Survival Curves 3. Hazard Functions
LAB 6: Diagnostic and Screening Tests 1. Diagnostic and Screening tests 2. Sensitivity, Specificity, and Accuracy 3. ROC curves 4. AUC calculations 5. Cut-off values
LAB 7: Discrete Distributions 1. The Binomial Distribution a. Calculating Probability and Percentiles 2. The Poisson Distribution a. Calculating Probability and Percentiles
LAB 8: Normal Distribution 1. Calculating Probability and Percentiles
LAB 9: Random Sampling and Central Limit Theorem 1. Review of distributions a. Setting to show that they can be approximated by normal distribution 2. Random sampling a. So that getting good estimates with random samples does not require very large samples in many cases 3. Central Limit Theorem a. Apply it to the data in the exercises and see if resulting distributions are normal
LAB 10: One Sample T Methods 1. 1-sample t-tests a. Hypothesis testing and Confidence Intervals 2. Making new variables in R a. Using a variable in the data set to make a new one
LAB 11: Paired t Test 1. Paired t-tests in R a. Hypothesis tests and Confidence Intervals 2. Visualizing paired data in R
LAB 12: Two Sample t-test 1. Two-sample t-test in R a. Hypothesis tests and Confidence Intervals 2. Violin Plots in R
LAB 13: Hypothesis Tests and Confidence Intervals for Proportions 1. One Proportion tests in R a. Hypothesis tests and Confidence Intervals 2. Two Proportion tests in R a. Hypothesis tests and Confidence Intervals
There are several ways iblir can be used to teach the labs. Each is briefly described below.
This way of using iblir is described above. Simply install and load the package. The greeting message will tell you how to load each lab. The Package website also provides this information.
This assumes you know how to set up projects in Posit Cloud. This is not very hard.
With this approach, students run R and
RStudio in the cloud through their browser, and don’t have
to install anything on their local computer.
For this approach, set up a new PositCloud project and install the
iblir package (which will also install a ton of others).
Then give students the link to the workspace. They log in, load the
package and can use it as if they were running it locally.
Note that a free RStudio Cloud account might not be enough and you might need to upgrade to a paid account.
This assumes you have a shinyapps.io account and now how to deploy apps. This is not too complicated.
Follow the shinyapps.io
user guide to set up your account and configure
rsconnect.
Next, load iblir on your local computer and run the
deploy_labs() function, which will send each lab to your
Shiny server. By default, all labs are deployed, but you can specify
which ones by providing the function with a vector of lab numbers, e.g.
deploy_labs(labvec = c(2,5)) will deploy only labs 2 and 5.
The labs will be on shinyapps.io at URLs like this one :
https://YOURACCOUNTNAME.shinyapps.io/lab1/ etc.
This assumes you have your own Shiny server up and running. This requires a good bit of technical expertise. One option to run your own Shiny server is to use DigitalOcean and e.g. the analogsea package.
Load iblir, then run the get_labs()
function. It will open a file browser and point you to a zip file with
all the labs. Copy that zip file to your shiny server, ideally into its
own folder (e.g. a folder called iblir). Then extract. You
should end up with a folder for each lab,
e.g. /srv/shiny-server/iblir/Lab1,
/srv/shiny-server/iblir/Lab2, and so on. Make sure the
iblir package is installed on your server. Also, make sure
permissions for the folders are set correctly.
Once all is set up, students should be able to go to the Server/Lab
URL to access each lab, e.g., by going to
shinyserver.myplace.edu/iblir/Lab1 and so on.
If you want to make changes to the labs, here is some brief information to get you started.
If you enhance or add to the labs, we would love to receive
your contributions and make them part of iblir! For more see
Contributing to the Package below.
The best way is if you edit the whole package.
First, to make sure you have all needed package dependencies, install iblir from GitHub as described above (no need to load it).
Now, fork iblir on GitHub, then clone it to your
local machine. Open the package repository by opening
iblir.Rproj with RStudio.
All labs are within the /inst/tutorials/ folder of the
package repository. Each consists of a main R Markdown
(.Rmd) file, and as needed a folder images for
images and a folder www for other files. If you want to
modify an existing lab, open the corresponding .Rmd file
and edit it. If you want to add a new lab, start a new folder and follow
the content and structure of the existing labs to create a new one. Keep
the naming consistent (i.e., call it LabN with
N whatever the next number is).
As you make changes, run Install from the
Build tab in the upper right pane of RStudio to rebuild and
re-install your updated package. Test it to make sure everything works.
You can also run the Check command to make sure the package
is fine (you can ignore notes). If you use some new R
packages for your lab, you need to open the DESCRIPTION
file in the main folder and add the new package to the
Imports list.
Once everything is ready and works, take all the labs, put them in a
zip file, and replace the existing alllabs.zip inside the
/inst/allabs/ folder with your zip file.
Finally, push your updated package to GitHub.
Students/teachers can now use your version of the package in exactly
the same ways as described above, with the only difference that any
install command should now point to your fork of the package (e.g.,
install via install_github('yourgithubaccount/iblir')).
As a last step, please contribute your enhancements to the packag! See below.
Again, we recommend using the approach just described. But if for some reason you are not comfortable editing the package, here is another way to make adjustments.
Start by loading iblir and running the
get_labs() function.
Unzip the folder, find the lab you want to change, and make any changes you like. You can also add a new lab as described in the prior section. Test your changes to make sure everything works.
Once all works, you can then copy your new labs to your own shiny
server as described above. If you want to deploy to shinyapps.io, call
the deploy_labs() function with an additional input
specifying the folder in which all your new labs are stored, e.g.,
deploy_labs(labfolder =C:’)`.
This only works if you deploy the labs to a Shiny server. For local use or through Posit Cloud, you need to use the previous method.
If you made changes/improvements to the package as described above, please send a pull request. We would love to get contributions from others.
If you plan on making large changes, it might be best to file an issue first and let us know. Then we can discuss and make sure your planned modifications can be seamlessly integrated.
If you don’t feel comfortable editing the package, but have some other feedback, e.g., bug reports, feature requests, please file a GitHub issue. Any contribution is very welcome.
See the package website for useful additional information.
The main content of this package was developed by Brian McKay. The package is maintained by Brian McKay and Andreas Handel. A full list of contributors and a Bibtex entry for the citation can be found here.