Course: Spatial Analysis in R

Course title: Spatial Analysis of Ecological Data in R, organised by PR~statistics.

The course was really extensive. Each day we did at least one new "module". Have a look at their website for the list: Spatial analysis course.

It was one of the hardest courses, if not the hardest one, I have ever done. By Wednesday afternoon I was nearly in tears. Frustration with R was taking over - the course material was really tough and pretty much all new to me and on top of that we had to write our R scripts from scratch. Which for me was as close to impossible as one can get. While I understood the concepts I really struggled to apply the knowledge effectively and work through assigned tasks, as my coding knowledge was letting me down.

I have to admit, during the course I wasn't sure how much use it all would be, since I felt like I wasn't getting through enough of the practicals. However (and this is a big however!) during the last day we split into smaller groups and discussed our own data. I managed to get some feedback on my ideas and some tips about future analyses. It was really, really, really useful! In the end we were also given copies of all the code necessary to complete all the practical tasks. I really think that I should be able to go over things again in my own time and use everything I learned.

This course ate more than half my yearly PhD-expenses budget, but I think it was worth it in the end and I would recommend it if you are interested in spatial analysis. Things would probably be easier if you know a bit of R (even better if you know a lot of R). If you don't know R, but you want to do it, be prepared for some really hard work and likely a bit of confusion and frustration too.

There is something more this course reminded me of, which I think is really important: if you struggle, talk to other people! I can't stress this enough. So many others on the course struggled and knowing that made me feel less alone and less hopeless.

Course: Quantifying the environment

The full title of the course was "Core and advanced environmental statistics training: Quantifying the environment – Part II" at the University of Glasgow. As a NERC funded student I could attend the course for free - thank you NERC!

It was a week long course, it was rather intense and it covered three areas: flexible regression methods, spatio-temporal modelling and functional data analysis (FDA). I was particularly interested in the spatial modelling as I might do a chapter on it for my PhD.

The course was very much aimed at statisticians and quite a bit of previous knowledge was assumed. For instance, on one of the days we had a set of lectures which were a condensed version of a 20 h long course for people doing Masters in statistics. As I have no real background in stats I found a lot of it rather hard. Keep in mind that I haven't attended Part I, maybe things would have been easier if I did.

What I found a bit suprising is that a lot of things we were doing were "at the edge of statistical knowledge and research". For some reason I have expected the field of statistics to be more advanced. This might sound silly, I guess, but I just never thought deeper about where our statistical knowledge is at. Now I know! It seems that some things I was planning to do might be harder than I thought, yet the majority of them should be doable, which is great. There were a couple of things that are not-so-doable yet, mainly due to lack of packages available - the knowledge is there, but there is no easy way to utilise it by an average non-statistician.

Yet despite the fact that the stats were hard, I found the course really informative. I think this was due to a clever delivery strategy. The lectures were clear and accompanies by practical sessions. All the R scripts for the practicals were provided. This way I could work through the material and focus on trying to understand it, instead of fighting with R. I think if I had to come up with R scripts by myself I would have been completely lost. Additionally, I have all the scripts for future reference and I can be sure that the code is decent and without errors.

All in all I'd recommend the course, but be prepared for a long week and quite a bit of work!