Principal Component Analysis (PCA) in R (presence-absence data)

In this tutorial, we discuss what a principal component analysis (PCA) is, walk through an example in R using species presence-absence data, and create and interpret a PCA biplot.

[lobuziak]

BY FAR THE BEST AND MOST USEFUL VIDEO I HAVE SEEN SINCE BEING THANK YOU FOR EXPLAINING THIS TO ME!

[gijs]

By far the best video I could find on this topic, thanks a lot!

[mollyharrism]

OMG finally an explanation of this that makes sense to me. Thanks!

[akritiashesh]

I find this tutorial really helpful. Thankyou for making this video.

[NeverLoudxx]

Oh my god I'm so happy I found this video before my ecology final tomorrow. Thanks so much 😭💖

[kelseyhoppes]

Video was incredibly helpful in walking through and understanding PCA

[petersantana]

Incredible video! Simple and straight to the point! I wish you well. Regards from Brazil 😇

[federelix]

Thanks so much for the video, helped me a lot, this was exactly what I needed for my ciliate data! Greetings from Austria:)

[ancamihaelasuteu]

Hello :) first of all, I would like to thank you for THE most straight forward and on the subject of ecology/biology video I have ever found!! It helped me understand and solve several mistakes and questions I made/had. REALLY, thank you! I would appreciate and love if you could find the time to do more videos like this.

I only have one question: how to avoid the overlapping of the site labels/species labels? I have 30 sites and more thank 30 species, the problem being that I can not see what sites are overlapping :(

Thank you again and I wish you success and great accomplishments in your field of study!

[natalialopez]

Great explanation! Can I do a pca using rda with numeric variables of 11 levels?

[sodayaummi]

This video is really helpful to me, and i would like to know about the variable, is there any minimum amount for the variable in using PCA? such as 5 variables of places with 6 or 7 parameters, could it be use PCA to solve it? Thanks

[PedroAndrésMuñozSantibáñez]

Hello! I just wanted to thank you for this amazing video! I also would like to ask you something. What would you do if you want to represent your species under different habitats (e.g. Forest, Meadow and Scrubland) but each habitat has their own amount of sites (e.g. 20 sites each one)? Would you combine your sites to represent the habitats? I'm a little bit lost about which approach I should take to work with my data. Thank you so much!

[slimestage]

Great video! Super helpful!!
Would a PCA like this, for presence absence of species, be able to include explanatory variables that explain the distribution of the community in the PCA? Your video was super helpful and I was able to run the PCA with my data but now I'd like to visualize in the PCA plane how other variables relate to the spread and layout of the species in the pca. Is a PCA still even what I should be using?

[ebd_]

Hi, thanks for the video!!! really helpful and helped me through the first step.
The species data I have is also species/absence. However, because I'm dealing with plants, I have p/a data for 159 species :( and therefore my explained variance values are very low (i.e. like 0.03 for PC1). What would you recommend in this case? Should I take out less important or very rare species?

[liamtaylor]

Epic channel so far! Thanks for the video, from South Africa :) Could you do a PCA video with some environmental variables linked with the species' absence-presence? Would that be a scaling = 2 PCA?

[grego]

Yes but how do the principal components that come back when you get the summary from a PCA in R correlate back to the variables you input? I have yet to make sense out of this, nobody seems to explain it clearly and simply, not my teacher and not 1 single youtube video I've watched. I'm lost...

[alandent]

I guess by now that you've noticed that your spelling of "principal" is incorrect.