This is good. I have not had a chance to play around with Elastic Search, so I will definitely look into that. We actually have Mark's Solr stats running now. What it lacks‹really, what they all lack‹is a sure-fire way making only counting real users.
If you are anything like us, then you know most of your traffic is robots. You obviously don't want them counted in your stats. The blacklist approach does not work; it doesn't come close, unless you have a full time person working to keep the list up to date. The Google Analytics approach is better, but it comes with its own set of problems. Like, links from Google have to go to the item page, not directly to the PDF, in order to be counted. A third idea I've been kicking around is to augment the current system with ping-back, so only the downloads and page visits that respond to the ping back are counted. That would probably take the least work to implement, but it could have issues as well. This is going a bit long, but let me add that I am surprised that this is not a bigger concern within the community. Accurate statistics are key to measuring an article's impact, and the impact of the repository itself. It definitely influences buy-in: both faculty participation and institutional support. Anyway, thanks for the input. I wonder, does anyone else have any insight into gathering accurate statistics? Cheers, Bill On 10/24/12 6:44 PM, "helix84" <[email protected]> wrote: >Hi all, > >if you want to collaborate on this, it would be best to create an >issue (or multiple issues) in Jira describing what needs to be done >and who's doing it. It will prevent duplication of work and allow >everyone to check the status of such work. It's also desirable to link >to code on GitHub from the issue(s). > >You may also want to coordinate with Peter Dietz (should be on this >list), who has developed another statistics implementation (Elastic >search) in addition to Solr statistics (Mark Diggory will want to have >a say in that, too). It may turn out that there's a need for a >framework supporting different statistics implementations/backends >(Solr, Elastic, Analytics, ...), while allowing to use the same >presentation elements (like graphs) with any of them. > >Regards, >~~helix84 ------------------------------------------------------------------------------ Everyone hates slow websites. So do we. Make your web apps faster with AppDynamics Download AppDynamics Lite for free today: http://p.sf.net/sfu/appdyn_sfd2d_oct _______________________________________________ DSpace-tech mailing list [email protected] https://lists.sourceforge.net/lists/listinfo/dspace-tech
