i've hit a bit of a tricky visualisation issue;

I am quantifying some land use and land use change (LULUC) over a number of years. Typically, for data going from year to the next, a land change matrix is pretty handy;

enter image description here

In the above image you can see what percentage of the total land present in 2005 (the left column) becomes/stays as what in 2006 (the row along the top). Eg category A staying as category A is 50.2% of all land in the dataset, category B changing to category F is 2.4% of the land etc.

  • it is important to realise that this represents quantities of change from 05 to 06, not absolute quantities of that land category in either 2005 or 2006.

However i have a dozen change matrices, all important, but i'm slightly loathe to show them all in my results.

  • I thought this multivariate time series idea for R was pretty nice, but it will only show absolute quantity from one category to another.

  • A lattice/trellis plot where by the land categories are the 'outer' axes forming the structure of the trellis while 49 bar charts showing % change over each year step would be displayed, but this sounds horrid.

  • A 3D plot where by the matrix above is in a 3D plot like a set of tiles, the colour of the square is used as the quantity indicator (think heatmaps and geom_tile in R), and then the data in the time series is stacked on top of it, almost like a stacked bar chart but each bar would consist of 'tiles' reaching the same height but colour to denote quantity at each level. Sounds complicated.

Are there any good ideas out there for this issue? perhaps multiple matrices are the clearest way.

(p.s. R is preferable but anything that works is fine too :) )

  • i'm working on using Sankey Diagrams – Sam Jan 19 '16 at 9:24

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Browse other questions tagged or ask your own question.