# What type of PCA to run with only two environmental variables and 1 factor?

I study soil insects, and sample monthly for insects. Each month, I sample at 8 different sites. Each site is divided roughly into 4 meter square quadrants. From each quadrant, I pick out a random sub-sample soil and look for insect abundance. I repeat this each month. I also have temperature and moisture measurements for each sub-sample. Of the 8 sites, 2 are in the woodland, 2 in urban, 2 in rural, 2 in industry.

site   setting   quadrant  species   abundance   temperature   moisture

1     Woodland      1       A          80          6.2           32
1     Woodland      2       B          2           5.2           12
1     Woodland      3       C          0           6             40
3     Urban         1       A          3           5             20


...

I want to know if insect species separate out on these axes (setting, moisture, temperature). I'm wondering what is the best type of PCA to run for this analysis? PCA appears to be for quantiative analysis whereas as PCoA for qualitative..but this maybe a mix of both. I'm a PCA newbie, so please explain in detail.

• Why do you want to use PCA? – Michael M Apr 2 '20 at 19:03
• Why not use PCA? I'm also considering a mixed effects model..but I find the PCA visualization much more easier – Biotechgeek Apr 2 '20 at 19:06
• PCA is usually used to summarize many quantitative variables into few orthogonal axes. Here you only have 2 quantitative explanatory variables and one factor to explain one quantitative response variable. If you're looking for a visualization tool, bubble plots might be usefull? See: r-graph-gallery.com/bubble-chart – Circus pygargus Apr 9 '20 at 16:43