Seb, I have now applied the patch given in https://github.com/eprints/irstats2/issues/5 and started “process_stats” again. The results now look more reasonable than before. Thank you very much for the good work.

Volker

Hi Seb,

 

we also installed IRStats2 Beta at http://epub.ub.uni-muenchen.de/cgi/stats/report/. Here are my remarks:

 

1.     I am wondering about the results. Looking at the report you see “Gottfried Keller: Kleider machen Leute” (Author: Rolf Selbmann) is one of the most downloaded publications. But the author should also occur in the “Top author” list. But his name cannot be found in that list. The set is defined as follows:

       {

                    'name' => 'authors',

                    'field' => 'creators_name',

       },

       because most of the creators have no populated ID-field.

I'd be interested in debugging that issue with you - I'll email you outside the mailing-list if that's OK.

2.     If you go to “Filter items” you will get a lot of author names several times. I suppose it depends on the number of the publication the author has published in the archive. It makes it nearly impossible to get all the downloads of one author in one report.

Ho you just found a nice bug in the name extraction, see https://github.com/eprints/irstats2/issues/5

I'll have this fixed quickly on github and will update the EPM.

3.     Just cosmetic: If a name contains a hyphen like “Hans-ulrich” the name part after the hyphen is changed to lower case.

OK I'll flag this as a bug for now and will see how I can improve the rendering of double-barrelled names ( https://github.com/eprints/irstats2/issues/6)

Thanks for your feedback!

Seb