Just to sum up, I do understand what the point is, but I don't think we're talking about framework level functionality here. A plugin (application in web2py terms) specialized for this, maybe, but I feel it's not the job of a web framework to do user tracking, data mining and similar high level tasks. Special care must be taken for such an addition not to incur overhead on applications that do not need it. I do support modifications that could make it easier to match a functional presentation of the system to actual web requests (for example form.accepts can play tricks on page trackers).
As for google analytics (GA) and related services, do not underestimate it. You can get a lot, really really lot of info out of it. Many 3rd party commercial applications provide similar functionality building on the fact that most people are not familiar with advanced GA usage or how to retrieve raw data from GA. On Mar 2, 6:01 pm, Francisco Gama <[email protected]> wrote: > > > Notice that is still somehow incipient compared to what I had in mind. > > > In your slideshow (second pic) you see how many people have given up, > > > and where, but you don't see why. > > > Can you determine which was the cause? what is triggering it? > > > No matter what the tool, this will always be a task for an analyst. > > You can do a lot of data mining, but unless for predefined use-cases, > > the conclusion will have be made by a person. > > Yes, but persons need data too to make their own conclusions and I > don't see why should we neglect predefined use-cases since they are > very good resource to get useful data. There are easy ways to do this > based of cookie collection (there are passive tools for this as well > as firefox extensions) that give you this even in a blackbox stand (no > need to open the source...). > > To think about: > If I just know that someone with IP 120.193.192.12 had a non- > applicational problem at 15:11 GMT +00 in my /index.html . > With google analytics you can also see that more 1000 users had a > problem there in little time. > > But Can you see if it is being triggered by the browser? or is it the > agent's language? is it a CSS issue? I don't think you can determine > it. Many times the only way to know the problem is to actually check > the same page the user got and correlate it manually with the error > reported but for this you need a copy of the page error. You can't see > it with Google Analytics. > > This is not something you can't put into a medium or large size > application, all you need is to do a wise solution just as the tool I > saw in the presentation did. Think of Data clustering and > fingerprinting. > > > > > > Can you somehow correlate it with the browser/agent (to know if it is > > > a browser specific problem)? > > > Can you correlate it with time? > > > Actually, yes. Google calls this 'advanced segment'. You can create > > and cross reference almost any field from google analytics with and/or > > statement with most of the reports (either an existing report, or a > > 'custom report'). So you can check for example how many Spanish > > speaking iPhone users from Los Angeles who have already visited your > > site more than 10 times failed to complete a form after 10pm. Granted, > > not all reports support advanced segments yet, but every now and then > > another pops up. --~--~---------~--~----~------------~-------~--~----~ You received this message because you are subscribed to the Google Groups "web2py Web Framework" group. To post to this group, send email to [email protected] To unsubscribe from this group, send email to [email protected] For more options, visit this group at http://groups.google.com/group/web2py?hl=en -~----------~----~----~----~------~----~------~--~---
