Currently I use PcORD Version 5. I am interested on PRIMER 6 now but keep looking for other options.
Primer 6 with PERMANOVA + add-on is very user-friendly and useful for community analysis.
'R' is also incredibly powerful and free. However, it is not very user-friendly and will require a lot of learning.
Cheers.
I like Primer 6 very much. It has lots of options and is easy to use with its windows based interface,
I heard from colleagues that MatLab is a very powerful program (I have never used it). But, just like R, it seems to require lots of effort to master the program.
The various packages written for R (vegan etc) are very capable and allow analyses that cannot be undertaken using other software. There is a fairly steep initial learning curve, but provided that a little time is allowed for initial learning is practical and powerful (and cost-free).
It depends on which analyses you want to perform. I'm using PC-ORD 6 in combination with SPSS 22.0 for ordination, cluster analysis, correlation coefficients, ANOVA, Kruskal-Wallis, ... . For other analyses like linear regression, BRT, functional and phylogenetic analysis, I use R. I think it is very useful to learn how to work with R. It is a very powerful program and not that hard to learn.
You can also have a look at CANOCO (http://www.canoco5.com/) - I worked with version 4.5 and its a fine piece of software.
CANOCO 5 is really very useful software, with its reference books. I also like PAST which has lots of functions, but does not provide strong graphical capabilities. Unfortunately there is no single userfrendly program doing everything. If you want to do some advanced regression type analyses, there is very nice program - BRODGAR (http://brodgar.com/) with reasonable price. To my knowledge it is only userfrendly program doing almost all kind of regression kind analyses as well as time series and most multivariate ordinations.
With regards of all suggestions, its seem that every program has their own advantages:
PCORD 6, R, Primer 6, CANOCO.
My colleagues from Texas A & M have suggested PCORD and it worth $199 (student price) that Im using now. Do you thinks we need to master more than 1 program ?
I think you should. You definitely need to learn a standard statistical program like SPSS, statistica or SAS, because programs like PC-ORD and CANOCO don't have all the features you need. Of course, R can do almost everything...
Hi. As other coleagues here I would strongly recommend PAST, it's free and quite use friendly. It is very powerful and has a lot of options for diversity analysis, ordination and classification, as well as other multivariate and parametric tests.
R is also great because it can do pretty much anything, if you download the right packages, but it's not user friendly. However there are several multi-matrices analyses that are now currently carried out in community ecology that almost only can be performed with R.
I find SPSS very helpful too. Yes, it would be best for everyone working with communities to master more than one program.
Good luck!
I used Primer for analysis of bird communities - very useful and user-friendly
Fifth version of CANOCO is very handy, especially for multivariate stuff. Many new features have been incorporated to make it more user-friendly than CAN4.5.
Dear Azwandi Ahmad,
I agree with Carvallo and Kuber Bhatta. In community ecology for gradient analysis, different packages of R (such as Vegan) or CANOCO either version 4.5 or version 5.0 are very useful. Among these, I perfer R because it is free and easy for aanalysis.
I use CANOCO or PRIMER, with the program dependent on the data. If you have gradients in assemblages responding to underlying gradients in environmental variables, then the analyses in CANOCO are best. If your underlying environmental variables are categorical (e.g., before/after, high/medium/low), then the analyses in PRIMER are best.
Primer 7 is now available, with many new techniques.
You should know what you want to do (and why), and choose the software to do it accordingly, not the other way around. Primer has an underlying philosophy based on a non-parametric approach which is widely applicable.
R is fine, but be aware that many of the routines/libraries are not written by, or checked by, statisticians, as far as I am aware.
- Badges
- Science topic
- Similar topics
- Computer Science and Engineering
- Software