{"id":521,"date":"2010-07-08T16:36:33","date_gmt":"2010-07-08T15:36:33","guid":{"rendered":"http:\/\/maurice.vanderfeesten.name\/blog\/?p=521"},"modified":"2010-07-08T16:36:33","modified_gmt":"2010-07-08T15:36:33","slug":"or2010-dspace-1-6-usage-statistics","status":"publish","type":"post","link":"http:\/\/maurice.vanderfeesten.name\/blog\/or2010-dspace-1-6-usage-statistics\/","title":{"rendered":"OR2010 – Dspace 1.6 – Usage statistics"},"content":{"rendered":"

This slideshow could not be started. Try refreshing the page or viewing it in another browser.<\/p>

<\/div>
<\/div>
<\/div><\/section>\n

\"\"<\/a><\/p>\n

@mire has build a GUI<\/a> for statistics on Dspace at the administrator side.<\/p>\n

User requirments: a survey has been held. Repository managers would like to see the popularity of items, collectiions or communities.<\/p>\n

Technology<\/strong>: They use Apache Solr<\/a> (fork of Lucene<\/a>) for indexing and fast search and faceted browsing.<\/p>\n

Advantages<\/strong>: statistics are context aware. This means the items are aware of the D-space hierarchy, group items, bitstream and pageviews<\/p>\n

Visualize<\/strong>: using the SOLR faceted search<\/a>, the popular items, collections and communities can be easily shown in matter of milliseconds. This can be represented in charts, and data tables.<\/p>\n

Reporting: <\/strong>Also an reporting function is available where the administrator can create reports with the facets they like by the press of a button. This can be exported in different wordprocessing formats and dataformats.<\/p>\n

thoughts and possibilities<\/h2>\n

Within Knowledge Exchange “Usage Statistics” strand the country initiatives PIRUS (UK), OAstatistik (DE) and SURE (NL) are working together to create Guidelines to facilitate the exchange of Usage Statistics<\/a>. These statistics come from distributed repositories and will be stored in a central database. This database will provide an API<\/a> where the statistics data can be used in GUI services.<\/p>\n

OpenAIRE is creating a GUI with the datavisualisation API’s from Google<\/a>. (Like we have seen in the presentation about Usage statistics and Google Analytics<\/a>) However, according to Ben Bosman it is very risky to rely on third party visualisations you cannot control. API specifications can change and the visual design can change.<\/p>\n

I was thinking about using the GUI from @mire. In theory it could work, becouse their internal format for storing usage events look very similar to the OpenURL Context Objects. However Ben was very reserved when this idea was postulated. He told me that their GUI is tightly integrated with the Dspace\u00a0 framework. The eye-candy of their module is great, but it is not OpenSource<\/a>. This might not be a deadend. We could discover some posibilities.<\/p>\n

What is needed to create a GUI with either OpenAIRE or @mire<\/h3>\n
    \n
  1. User requirements study: what type of person\/role needs what type of information.<\/li>\n
  2. @mire GUI: A repository tab in the administrator GUI<\/li>\n
  3. @mire GUI: the xml format they are using for usage events needs to be mapped to the KE format for OpenURL<\/a> Context Objects. It is quite similar so an XSL might do the trick.<\/li>\n<\/ol>\n

    Tip for the people who are implementing OAI-PMH<\/a> statistics services on repositories. (Dspace, Fedora, eprint)<\/strong><\/p>\n

    If you want transmit the Geo location of countries or Cities with the OpenURL Context Objects before scrambling the ip-address, you can use GeoLite<\/a>. A free version of the mapping of ip-addresses to countries and cities. An API<\/a> implementation can be used.<\/p>\n

    \"Enhanced<\/a>