(A guest post by Camiel de Koning)
————–
When trying to replicate, verify or extend empirical research of others, a researcher generally encounters many time-consuming barriers and there are often many prerequisites. Fittestmodel has the objective to overcome many of these problems, by presenting a webapplication that allows users to:
- use but not having to install R.
- quickly incorporate datasets from a massive data library.
- post their analyses instantly to the included messageboard.
- extend/verify/modify others’s analysis easily.
- find the context and source of used datasets immediately.
- let others to help solve your problem.
The engine behind Fittestmodel is R, which is accessed through the nice Rserve interface. The user can enter the normal R commands, however the commands that interact with the server (like ‘system(‘format c:’) are filtered out (it would be nice if a future version of R could include a kind of firewall for the host system). Fittestmodel also includes a time limit imposed on the calculation to prevent users to lock the application indefinitely (the price of scalability is a set of limitations).
You can get a nice overview from the various youtube Demo’s and the FAQ on the www.fittestmodel.com
Here is one example:

Zero Inflated Models and Generalized Linear Mixed Models with R.
Zuur, Saveliev, Ieno (2012).