Re: CAT - tips
Carlos Carroll
Manuela - Thanks for your helpful note. I hope you don't mind, I am cc'ing the Google Group on this reply as well as this info may help other users.
We have encountered this problem with "locale" settings previously.
We unfortunately did not implement code necessary to make the CAT "locale-aware", so as you suggest we just need to emphasize in the manual that this is a problem for non-Anglophone locale users and that there are workarounds.
On you point 2, one issue is that some text editors do not recognize the end-of-line in the header, so it appears in your left image that the first line of values is appended to the header.
There may be additional issues with the default settings for csv import in ArcGIS on European locale PCs. I find ArcGIS's menu-based join a pain to use and actually do this file import in Arc Workstation.
Please contact me again if you have questions regarding the reindeer analysis.
I plan to post a new version of CAT as soon as Networkx version 1.8 is completed (within the next 1-2 months).
Dear developers of CAT,
First, thanks a lot for this fantastic tool, which I am currently exploring. We are currently writing a paper for a Special Issue on animal movements in Journal of Animal Ecology, in which we review theoretical bases and methodological approaches to identify movement corridors, and compare the performance and ecological significance of different methodological approaches. Hence, I definitively intend to use your tool, and apply it to several friction maps based on a large dataset of GPS-monitored wild reindeer in Norway. For the time being, I only run the examples you provided, and I encountered a few issue which I report below:
1) I work in Norway, and the local (I guess European) default pc settings use a comma as separator for decimal numbers. This implies that the text files produced by my pc using CAT have commas to separate decimals, and this would cause CAT to show error messages when running the analyses. To avoid error messages, at first I simply transformed commas into points using a text editor, but this would only allow solving a few of the problems - e.g. in exercise 2 I could run the Connectivity tab without getting error messages. However, I encountered other problems later – e.g. in Exercise 3 when I press “run” in the Connectivity tab (both in the first and second analysis) I get the following error message:
I figured out that that the latter problem was due to the fact that the probabilty value (e.g. 0.02) in the CAT mask was by default written using a comma, and changing manually the comma into a point was not possible (CAT would automatically change it back to a comma).
Hence, the only way I caould make CAT work on my pc was to change the local settings of my pc, by specifying that decimals should be separated by points instead of commas. As I think that this problem is common to all European-based researchers, I would recommend you to underline this in your tutorial.
2) Likely due to some other format issue, the “join” function in ArcGis does not work directly with the text files produced by CAT. Actually in the “Join” menu box that appears - menu item #2 -, it is possible to browse the desired text file, but such text file does not actually “stay” in #2. This is likely due to some issue with the text file previously produced by CAT, which looks like the one on the left, below. To solve the problem, I opened such text file in excel, define commas as column separators, and save the txt file again. The corrected version (on the right, below) can be used in the Join function in ArcGis. For some reason unknown to me however, sometime ArcGis does not read the .txt file until its name is changed. Is is possible to solve these issues in a more elegant way in CAT?
3) Minor issue: in Exercise 3, Graphs tab., you write: “Examine the file hab5cond.txt in a text editor. Note how the edge values are the reciprocal of those shown in the first edgelist file we created (habcost5dist.txt).” I could not find any “habcost5dist.txt” in any of the previous exercises, and I suspect you actually meant “hab5dist.txt”
4) Minor issue: in Exercise 4, 1) graph tab, line 9: I guess you mean “upper” instead of “lower”
I am now starting to work with our own reindeer dataset and – if possible - I may contact you again regarding this paper we’re writing. Also, if you’re interested, we may consider making our data avaialble for future versions of the CAT manual.
Compliments again for this useful tool (and for the very interesting paper in Cons Biol)!
Best wishes,
Manuela
________________________________________________________________
Manuela Panzacchi
Researcher
Norwegian Institute for Nature Research - NINA
Postal address: P.O.Box 5685 Sluppen, NO-7485 Trondheim, NORWAY
Delivery/Visiting address: Tungasletta 2, NO-7047 Trondheim, NORWAY
Phone: +47 73 80 14 00 • Cell: +47 90911097 • www.nina.no
________________________________________________________________
