We are working on a web based application (implemented in JAVA) on commodity prices and one part of it is interactive charting. I provide a simplified example here. We have a table in Mysql database where we have information on commodity prices in US states and counties. One aspect of the application is to create interactive plots based on user choice. For example, if the user needs to see the price density in Oregon and Linn county then she chooses it from the menu in a webpage and it is rendered on fly with accompanying quantile information in a table. As the user changes state and county these plots and table change on fly.For our computational need we are using R (and use rjava to integrate it to our web application) and I know that if interactivity is not an issue this is a piece of cake in ggplot2, but I am not aware of any interactive version of R graphics framework (like lattice, ggplot2). We are exploring google visualization API but I am not sure we can have the statistical power we need in some of the plots.Please help.
You can absolutely do what you describe by making a call to R after each change in your web application. That's what Jeroen Ooms's application does. Or have a look at Jaime Love's R-Node application for an example of an R webapp that uses RServe.
R does not have any complete "interactive web graphics package" at this point in time. You can look at the webvis package for R/Protovis integration, but this doesn't include any interactivity yet.
Maybe a Javascript solution like protovis or humblefinance, maybe in combination with the brew package, might do the trick?
I don't know if it is powerful enough to suit your needs, but the gWidgetsWWW package can be used to make interactive graphics quite easily. Some examples and link to their source can be found here:
http://www.math.csi.cuny.edu/gWidgetsWWW/run/ex-index
--John