Preliminary inspection and data transformation
|Top Previous Next|
Having opened the data sets the species and environmental data will be presented in grids. The raw species and environmental data is viewed by clicking on the Sample/Species Data and Environmental Data tabs respectively. The data sets can be changed and transformed from the working data tabs.
Changes made to the working data will not change the raw data.
We will not undertake any transformations on the Dune Meadow data set. This is because the numbers in both the environmental and species data sets are both with a single order of magnitude.
We can find out about the range in our data using the Summary tab. Below it can be seen that the largest maximum environmental variable value is 11.5 and the smallest is 1. A similar inspection should be undertaken on the species data.
It is often the case that species vary greatly in abundance, over possibly 5 or more orders of magnitude. When there is great variation in abundance consider making a square root transformation. A log transformation might also be useful but most species data sets include many zeros so you would need to use a log x+1 transformation and this can be misleading is there are many low abundance species.
During this preliminary stage you should also check for zero sum rows and columns in your data sets as these will likely cause division by zero problems
later. This is done using the Handling zeros radio button on the Working data tabs.
Having completed these initial inspections we must now Examine and Select Environmental Variables.