Summary and Setup

This is the website for Part I of the Multi-dimensional Biodiversity Data Workshop, first offered Summer 2023.

Open to Collaborate


The Rules of Life Engine Project is committed to the development of new modes of collaboration, engagement, and partnership with Indigenous peoples for the care and stewardship of past and future heritage collections.

Suggested schedule


Day 1

Time Topic
9-9:45 Introductions, setup, workshop overview
9:45-10:45 CARE principles
10:45-11:00 Coffee Break
11:00-12:30 Abundance
12:30-1:30 Lunch Break
1:30-2:45 Abundance
2:45-3:00 Coffee Break
3:00-4:30 Traits

Day 2

Time Topic
9-10:30 Phylogenetics
10:30-10:45 Coffee break
10:45-12:30 Population genetics
12:30-1:30 Lunch Break
1:30-2:30 Finding multi-dimensional data
2:30-2:45 Coffee Break
2:45-4:00 CARE principles in data archiving
4:00-4:30 Wrap-up and mid-workshop survey

Learner expectations


We expect learners to have some familiarity with R, including a few basic concepts:

  • what a function is and how to assign output of a function to an object, e.g. object <- function(x)
  • read in tabular data, e.g. read.csv() for comma-separated values files
  • bracket indexing, e.g. understanding that data[,3] returns the third column of the data dataframe.
  • install and load packages, e.g. install.packages() and library()

If you need a bit of catching up, no worries! The Software Data Carpentries has a nice Introduction to R and RStudio workshop availible online that will get you up to speed. The episodes up to and including “Subsetting Data” will get you where you need to be for this workshop.

Software Setup


Checklist

  1. For this workshop, we prepared a virtual image containing all the necessary software, packages and settings we will need. To clone this image into your machine, you first need to download and install Docker Desktop from this link. Make sure to choose the right version for your computer (Windows, Linux, MacOS Intel Chip or MacOS Apple Chip).
  2. Once installed, open Docker and use the search bar on the top to look for iovercast/mess. This is where the image is stored.
  3. Once the image is found (the name iovercast/mess will show up in the drop-down list), you need to choose the processor type for your computer. If you are using a M1/M2 MacBook, in the tag drop-down menu, choose mac-m1. For any other architecture (Windows, Linux and non-M1/M2 MacBooks), choose intel. Then click on Pull.
  4. Once the pull is done, click on the Images option in the menu to the left. You should be able to see the new iovercast/mess image listed there. To get it running, click on the Play icon under Actions.
  5. Make sure to add a port for the image. When prompted for Optional settings, click on the arrow to open the drop-down options and type 8787 in the Host port option. Then click on Run.
  6. In the new screen, click on the 8787:8787 link below the image name.
  7. A page for rstudio-server will open on your internet browser, prompting you for the username and password. Type rstudio for both username and password.
  8. If an RStudio screen opens on your brower, you are ready to start working.

Data Sets


Checklist

  1. Throughout the workshop, we will be working with some simulated datasets stored online. They will be downloaded as we go through each lesson, following URLs that will be provided in the moment.

Troubleshooting


Callout

If you have any difficulties with these installations prior to the workshop, please reach out to the instruction team at biodataworkshop2023@gmail.com and we will be delighted to help figure things out!