Metadata Usage Statistics

From NGDCWiki

Jump to: navigation, search

Web Server Log

This covers pages with /metadata/ and /nmmr/ URLs. These are combined into a single set of summary statistics, since both types of URLs are used for viewing and authoring metadata, and there is much interlinking between them.

The NGDC web site supports metadata authoring and distribution for metadata at NGDC, NCDC, NOS and NSIDC, and CLASS (NESDIS Products). Note that NODC has a completely separate website for their metadata.

Our metadata is also "harvested" (copied and transformed) into other sites, such as Geospatial One-Stop (GOS), GCMD, CLASS. We've put in requests to them for their usage statistics, as those could represent much wider usage of our metadata.

Image:200903metadata-usage-chart.png

  • Total Pages: All pages viewed; includes the NMMR, web pages about metadata in general, and individual metadata records
  • Record Views: Views of metadata records for individual datasets, either as HTML, FAQ, Text, or other transformation
  • Record Downloads: Downloads of metadata records as raw XML
  • Record Edits: Edits of metadata records -- an indication of metadata management activity within NGDC/NCDC/NOS
  • Unique Hosts: Unique IP addresses; a rough indicator of the number of different people who viewed pages this month.
  • *November was the transition to new web servers, pinnacle and zenith, and did not have reliable log files. Numbers may have been inflated by two incidents of out-of-control wget sweeps by British Antarctic Survey and a Russian WDC.


Usage Total Pages Record Views Record Downloads Record Edits Unique Hosts
Apr 2005 11,390 703 0 2,201 328
...
Apr 2006 50,443 32,158 0 4,337 4,048
...
Apr 2007 90,857 19,688 62,708 1,151 6,248
May 2007 112,751 38,791 63,562 1,913 7,593
Jun 2007 154,800 55,521 90,243 721 7,186
Jul 2007 193,286 68,203 116,713 965 6,886
Aug 2007 159,211 67,092 83,961 508 7,111
Sep 2007 105,682 46,723 51,271 669 5,988
Oct 2007 91,974 39,476 41,827 1,294 7113
Nov 2007 94,119 38,460 40,131 1,997 6,715
Dec 2007 111,664 57,162 40,348 1,520 5,714
Jan 2008 135,146 80,170 39,909 1,463 5,805
Feb 2008 135,238 77,857 35,307 1,511 5,587
Mar 2008 165,138 93,952 53,602 1,043 8,133
Apr 2008 363,648 291,014 61,569 1,978 7,935
May 2008 217,396 136,731 74,159 1,078 8,143
Jun 2008 154,101 105,679 46,378 556 6,486
Jul 2008 138,586 108,386 25,594 1,433 5,204
Aug 2008 136,654 110,554 21,244 656 4,507
Sep 2008 137,165 114,287 30,391 352 5,145
Oct 2008 228,463 145,617 69,663 3,097 5,118
Nov 2008* 160,985 106,195 40,797 2,605 4,133
Dec 2008 156,477 120,125 39,470 1,664 3,754
Jan 2009 135,014 111,451 33,025 1,754 4,023
Feb 2009 187,160 156,432 50,713 1,636 4,028
Mar 2009 165,678 147,173 72,895 123 4,727


Count definitions

  • Record Views = Request(fgdc.jsp + xml-to-text.jsp + viewRecord.do) for nmmr/ plus nmmrview/
  • Record Downloads = Directory(metadata/published/ - published/z39logs/)
  • Record Edits = Request(editSubmit.do)

Details

Web log reports produced with Analog 6.0 (combining both production servers):

Note that there was a transition from peak/spire to pinnacle/zenith in mid Nov, and logs were unreliable for weeks afterwards.

Mar 2009 web-based metadata usage
Feb 2009 web-based metadata usage
Jan 2009 web-based metadata usage
Dec 2008 web-based metadata usage
Nov 2008 web-based metadata usage
Oct 2008 web-based metadata usage
Sep 2008 web-based metadata usage
Aug 2008 web-based metadata usage
Jul 2008 web-based metadata usage
Jun 2008 web-based metadata usage
May 2008 web-based metadata usage
Apr 2008 web-based metadata usage
Mar 2008 web-based metadata usage
Feb 2008 web-based metadata usage
Jan 2008 web-based metadata usage

Dec 2007 web-based metadata usage
Nov 2007 web-based metadata usage
Oct 2007 web-based metadata usage
Sept 2007 web-based metadata usage
August 2007 web-based metadata usage
July 2007 web-based metadata usage
June 2007 web-based metadata usage
May 2007 web-based metadata usage
April 2007 web-based metadata usage

April 2006 web-based metadata usage

April 2005 web-based metadata usage

As a point of comparison, see the web log statistics for the entire NGDC web site.

An Analog configuration similar to the new "Corrected" reports at the above link was used, reporting only actual page hits, not images and other auxiliary files. Hits from the local machines (panther, ridge, harrylap, metadata1, peak, spire) were excluded, since they generate internal utility/housekeeping hits, such as LinkChecker, PageWatcher and WebInspect.

metadata1 logs]$ analog6 -G +g./published200809.cfg

NMMR Z39.50

Some of the NMMR record sets have been made available via the Z39.50 protocol. This table shows the weekly Z39.50 hits since 2/5/06.

The Z39.50 protocol is used by Geospatial One-Stop (GOS) to harvest our metadata, and by FGDC Clearinghouse to search our metadata.

NMMR Z39.50 server log statistics

Personal tools