search

datacarpentry / r-raster-vector-geospatial

Introduction to Geospatial Raster and Vector Data with R
https://datacarpentry.org/r-raster-vector-geospatial
Other
113 stars 110 forks source link

Add a few VERY TALL (40000m) pixels in the DEM as an example of bad data values #25

Open lwasser opened 9 years ago

mjones01 commented 9 years ago

I like the idea. Does this mean adding a section to remove bad data values? Acknowledge issue in min/max section and then deal with it in bad data section?

ErinBecker commented 6 years ago

The current lesson 1 includes a discussion of bad data, and has the learners explore the data to see if there are any bad values. There are none in the data set being used.

screen shot 2018-06-22 at 9 01 04 am

Should this issue be closed? Or do we still want to add some bad data values before lesson release? My preference is not to change the underlying data at this point, as it will require some extra clean-up. Pedagogically, I do think it's a good idea to have the learners find and correct or remove bad data points. Maybe we can have this added after the first publication? I'm proposing a "status:wait" label.

lwasser commented 6 years ago

yes. it would be easy to modify the data but there is no good solution. i think a bit of discussion about bad vs missing data and options toa ddress it (without actually fixing it) could also work. The key learning lesson here i think is just to

  1. look at your data - histograms help
  2. quality check and read the metadata to determine how or if the data were "cleaned" previously
bkmgit commented 3 years ago

Some of the data has missing values, and this appears in the plots, for example: NEON-DS-Airborne-Remote-Sensing/HARV/DTM/HARV_DTMhill_WGS84.tif which is visualized in https://datacarpentry.org/r-raster-vector-geospatial/03-raster-reproject-in-r/index.html perhaps more could be done to understand this.