Milestones
List view
### This version aims to address several issues in the filtering and spatial aggregation of samples leading to spatial shift, taxonomic bias, and uneven sampling effort. **Specifically, this update will:** * Address sampling effort bias (and therefore spatial clustering) in GBIF records by implementing spatial and/or temporal clustering followed by rarefaction. * Address plot revisits in sPlot by identifying them and taking only the mean plot value * Control for plot size when aggregating sPlot records * Improved filtering of observations/plots with low-accuracy geolocations
No due date•0/1 issues closed- No due date•2/12 issues closed
- No due date•0/5 issues closed
Version 2 will implement several new features: * Using input data at native resolution rather than in downsampled form, in order to better exploit its high-frequency patterns and texture * Adding further predictors that may be important to estimate functional traits, such as topography, hydrological features, etc. * Joint modelling of different traits to better capture their mutual dependencies and to exploit synergies from joint representation learning * Improved filtering of GBIF and sPlot data to improve issues with spatial shift and imprints of data density on predictions
No due date•1/4 issues closed- No due date•6/22 issues closed