We have a slack group now, because slack is the thing: the address is https://caperusergroup.slack.com/
Twitter is still going @CapeRUser
At UCT the Division of Epi and Biostats have started a biostatistical methods seminar series, it currently runs once a month, you can find information here: UCT seminar series
We will start live meetings again soon.
It’s been a bit busy – sorry for the few months of silence. We’ll be taking a break until mid-January, then we’ve got a great line up of people to talk to us about R in 2016. Yay! R!
Last year was great, if you have visitors or anyone else that would like to give a talk next year please just drop me an email.
Also, I still tweet useful things on the twitter account. Follow @CapeRUser
Jasper gave a great session outlining some of the tools that can be used to deal with spatial data in R – I learned a lot – like R can pretty much do everything and it’s very straight forward. Plus nice pictures + learn to parallelise your code if you are working with big/high resolution data.
The very last @RevolutionR t-shirts were given away (hint hint: we need more!) and for the first time we did a dual live broadcast (to the UK somewhere and via Google on Air) which means you can watch the entire thing here and you can access the talk materials on Rpubs here. Jasper also got some great feedback on twitter,
Stay tuned for next talk, I’m hoping to get one more meeting in before the end of the year, probably Stellenbosch side.
Title: A Primer for Spatial Data Analysis in R
Date: September 10, 2015
Time: 17:00 for 17:15
Location: UCT Club, Upper Campus
RSVP: here (last t-shirts from Revolution up for grabs)
Presenter: Jasper Slingsby
Fynbos Node, South African Environmental Observation Network
South Africa, Centre for Statistics in Ecology, Environment and Conservation
Abstract: Points, lines, polygons and rasters – R can handle them all. My aim for this session is to give you the basics required to teach yourself spatial data analysis in R. I’ll start by briefly covering CRAN Task Views and how to install them (using the ’Spatial’ task view as an example), followed by some pointers for useful DIY resources, and then work through a practical example exploring fire history layers (polygon), topographic (raster) and locality (point) data using the libraries ’rgdal’, ’raster’ and ’sp’. The example will cover setting and changing projections and extents, raster calculations, rasterizing polygons and extracting data with a few neat tricks along the way. Depending on how much time that takes (and how my week pans out) I may touch on spatial interpolation, parallelising your code, and creating cool animations :) If there’s anything you’d like to see covered, email me (jasper at saeon dot ac dot za) and I’ll see if I can include it (but note that I am not a helpdesk…).
Great talk by Robert on building flexible, pretty, graphics in R using the ggplot2 package. We had a nice turnout (I forgot to count…) and the slides are available now in the Cape R repo – join the mailing list if you want access to this.
[I also forgot to take pictures…]
Our next speaker will be Jasper Slingsby from SANBI who is going to talk about “something spatial” but I know he’s been working with lots of geographic & environmental data and complicated Bayesian models so we can ask him hard questions. This talk will take place in September at UCT.
Thanks for coming out, see you next time!
This will be a great talk, come support the CapeR group, drink a beer (or soda), and enter the draw for prizes from our sponsor RevolutionAnalytics. RSVP here.
Date/Time: July 30, 2015 17:00 for 17:15
Location: Most likely at the UCT Club, Upper Campus, UCT
Speaker: Robert Schlegel, UWC
Beautiful R: Creating the next generation of figures with ggplot2
Are you tired of using excel to create the figures for your publications/ thesis? No longer satisfied with that old map of SA you’ve been using for the last two years? Does your workflow involve more than one program? This evening we will see how to very easily remedy all three of these problems using “R” and the package “ggplot2”. With the use of this one central package, and a few others to manipulate our data for plotting, it is possible to remove all other graphical and statistical programs from your scientific workflow. Besides saving one’s time, the code used to produce the figures in this manner is easy to read and manipulate, allowing one to have an amazing combination of both reproducibility in research and control in the graphical output. By the end of this talk an “R” user will have a clear understanding of how one goes about using “ggplot2” to create anything from boxplots to maps of South Africa or the world.
ps. if we can consistently get 30+ attendees, Revolution will give us better prizes and more money…bring your friends.
We had such a great talk by Prof AJ Smit from the University of the Western Cape last week, dodging graduation, load shedding and a last minute realisation that one of us had the wrong date. The talk was so interesting I didn’t even take pictures. Basically Prof Smit gave us a run down of the process and workflow that his lab uses to work with (a lot) of data – they have more than 40 years of fine scale (both temporal and spatial) coastal ocean temperature data that they work with. They also make fantastic visualisations in R (stolen screenshot of webpage).
The lab uses a combination of version control (git in this case), R, and LaTex for reproducible analysis. Apparently there is a package on the way…. It’s very exciting to see the principals of open, reproducible research and great visualisation getting implemented – plus the data is cool.
Lab details can be found here: http://kelpsandthings.org/smit_lab/ and thanks to Prof Smit for the talk and to sponsor Revolution Analytics for the t-shirts. We all <3 R.