Open romunov opened 7 years ago
That can be fun and interesting topic. We can devote one Meetup session just to R and spatial data. Do you have in mind any specific topics, presentations or people willing to present ?
I know Krištof does a lot of work in R, perhaps he could share a thought or two. I can show what R, combined with web tools and leaflet, can bring to the table towards presenting of spatial data (showcasing the wolfexplorer) in a dashboard-like manner.
what can be done with R that cannot be achieved with Py? Not trying to go towards Py vs R though :)
While I don't have much experience with (geo)Python, my guess would be, that as of yet, modeling is a wee more advanced in R.
would be interesting to try to reproduce the same model using both. just curious. but of course, at the end of the day, whatever floats your boat.
What I'm saying is that I'm not sure you can fit a, say, generalized non-linear model in Python. You might, but as I've mentioned, I don't have a great overview of Python capabilities. From what I've seen, people are connecting to R and its lme4
/nlme
capabilities to tap into non-linear models.
Which is the way to go, but adds another layer to your app/analysis.
wolfexplorer seems like great project, I will added it to suggestions for future topics.
+1 for having R presented at GeoDev. Wolfexplorer looks nice and would be interesting to hear more about it. Is it possible to map "only" original data or can one plot results of statistical analysis too?
Right now, wolfexplorer is designed to plot data and display some metadata. This is useful because we have a lot of samples for wolves and you need an interactive environment to make sense of it all. If you have information about relationships (parents, siblings...), you can also use this functionality to further explore the kinship in spatio-temporal relationships.
maybe @neatcoder could show us what he wrote for his master thesis for the upcoming project of reassesing of earthquake hazard in Slovenia. he did some coding in python.
Hi! Why not. The scripts are pretty much basic though, as they were written to aid data processing and not to make a complete software package. Some scripts take ASCII file as an input and some SHAPE files.
I have heared about R but I haven't used it yet. I think it would be very interesting to see how R can be used in connection with web and spatial data. I would love to attend this meetup.
+1 for topics both on Geo R and Geo Py. I am mostly working with raster data (satellite images) and am using both R and Python.
As @romunov pointed out, R has an impressive set of tools for raster processing. One can process time series of rasters very intuitively (raster and sp packages are excellent) and quickly despite huge datasets.
At UL FGG we are teaching Python (one language is enough :-) and therefore almost all my recent work is in Python. And everything that can be done in R can be done in Python. While R still has some advantages, Python is faster for my purposes.
@KrOstir using python to process satellite images sounds very interesting too. Would you be willing to give a presentation on this? Format of presentation could be very similar as at FGG, maybe more lika a workshop? But we shall move in another topic...
I studied at UL FGG. In my time we learned VB 6.0 and VB.NET. From which year on did they started teaching Python? Thx for info.
@VVidmar We should be Python only by 2020 :-).
Python has been introduced a couple of years ago and now we are using it more and more. In my subjects it is the main tool since last year. It is my hope that students will be fluent in problem solving with it before finishing bachelor and they should be able to use it as an indispensible tool (for data and spatial analyses) at master studies and theses.
@gobaRules I would be willing to present my work with satellite images. Sometime in spring 2018. Workshop would be interesting but more challenging. Number of participants should be limited.
@KrOstir what is your criteria for selection of participants?
@VVidmar nothing special. The number should not be more than e.g. 20 to work effectively.
R has a lot of functionality that handles spatial data. I know people from FGG work a lot in it.