Open wes-brooks opened 3 years ago
Ideals:
People will be familiar with vector and raster data. Other types, maybe not so much, so I'd stick to those two. Is there something on the list that you think demonstrates the overall important concepts well? Maybe spatial regression? From that they can learn that spatial data violates the independent samples assumption of most regular stats, and it helps them address hypothesis testing. Kriging and other forms of interpolation don't really help with hypothesis testing.
Thanks for the comment! It's super helpful to know something about the background of the audience. I think interpolation usually comes before spatial regression, which makes sense because doing spatial regression will usually require estimating the parameters of a smoother as part of the process. Has the group previously talked about the interpolation tools that are built into QGIS?
We have had a previous workshop about interpolation, but I can't say that people will have gone to it. This is more of drop-in when they like the topic crowd vs. a learn sequentially together crowd. Basic interpolation is built in to most introductory GIS courses - I know LDA 150 teaches it - so you can assume they'll know what an interpolated surface is, but they might not know too much about the details of it.
This is a basic introduction, so the topic list should be broad and use only stable, widely-used packages.
Ideas (even a brief dip into each of these may be too much for a two-hour workshop):
Of these, I think "concepts" is clearly the most important. How is spatial statistics different from mapping, or GIS tools, or non-spatial statistics? There's a need to explain the different kinds of data, even if I don't get into the analyses you'd do with the different data types.
The maptime participants are clearly familiar with some kinds of spatial data. From what I've seen, this includes vector data (points) with associated metadata. Some folks have used rasters - I am not sure whether there is familiarity with other kinds of areal data (though everyone will grasp the concept of something like a choropleth). Point patterns are maybe more esoteric - though people who work with wildlife are probably familiar.