Page History

...

1. DSpace comes with a very basic web statistics package which you can use to gather information on page views (not the same as file downloads) and general activities performed within DSpace (e.g. searches performed, logins, errors and warnings). This statistics package will analyze and parse all of your old DSpace logs (in dspace/log) to generate HTML statistical reports. It consists of a set of Perl scripts all beginning with "stat-" in your dspace/bin/ directory. (Note: In order to utilize these scripts you must have Perl installed!)
2. There are two basic options in dspace.cfg which allow you to specify whether or not your final statistics reports should be publicly accessible (

   Code Block
   report.public

   ), and where the final statistical reports should be stored (

   Code Block
   report.dir

   ). It's very possible you may not want to change either of these. However, if you change either option, don't forget to restart Tomcat (See Quick Restart in Rebuild DSpace )
   • report.public = false
   • report.dir = ${dspace.dir}/reports/
3. The DSpace Statistics package comes with its own configuration file, dstat.cfg. This file has some settings which you must change (marked with an asterisk below), and a few that you may wish to review:
4. *

   Code Block
   dspace.log

   *- the full path of your dspace/log directory

   • Code Block
     general.summary

     - actions listed in the DSpace log file which you want to list in the "Overview" section. You don't need to change these unless needed.

   • Code Block
     exclude.word

     - stopwords to filter out of search terms in statistics

   • Code Block
     exclude.type

     - Lucene search index terms to filter out of statistics (corresponds to the search indicies, see (Modify + search + fields).

   • Code Block
     exclude.character

     - Lucene special characters to filter out of statistics

   • Code Block
     item.type

     - Item types to find statistics for. Corresponds to the form values (see Change + a + form + value ) defined for your

     Code Block
     dc.type

     field, or any metadata field with an element named

     Code Block
     type

     .

   • Code Block
     item.floor

     and

     Code Block
     search.floor

     - specify the number of minimum accesses necessary before an item or search term is listed in statistics

   • Code Block
     item.lookup

     - specifies the maximum number of items to list Author/Title information for in statistics (all other viewed items are listed by URL)

   • Code Block
     user.email

     - specifies whether to display user email information in login statistics. For privacy, this defaults to

     Code Block
     false

     (i.e. do not display email)
5. *

   Code Block
   host.name

   and

   Code Block
   host.url

   * - The name and URL of your DSpace install, which will be displayed at the top of the statistics page.
6. In addition, the DSpace Statistics package comes with a mapping file (dstat.map), which maps DSpace "actions" which appear in the log files (in dspace/log) into human readable text. So, if you wish to update the text which appears in the statistics reports, or change the language, you can edit this file to do so.
7. First, you must modify each of these scripts slightly based on your own DSpace installation. Look for the following section in each of the stat-* files:

   • Code Block
     # Details used

     Code Block
     ################################################

8. In that section, you will likely need to modify any variable which specifies the full path of a directory or file location (these paths will all start with

   Code Block
   /dspace/

   by default). In particular keep an eye out for these variables:

   • Code Block
     $dsrun

     - the full path of the dspace/bin/dsrun script
9. *

   Code Block
   $in_directory

   - the full path of the input directory (for stat-report-* scripts)

   • Code Block
     $out_directory

     - the full path of the statistical reports output directory. This must correspond to the directory listed in the

     Code Block
     report.dir

     option in your dspace.cfg (see above)!
   • To keep things simple, you may wish to specify that both the

     Code Block
     $in_directory

     and

     Code Block
     $out_directory

     be the same location (e.g. dspace/reports/).
   • You may also wish to modify the

     Code Block
     $start_year

     and

     Code Block
     $start_month

     variables in

     Code Block
     stat-initial

     and

     Code Block
     stat-report-initial

     . These should correspond to the year and month which you wish to start tracking statistics from.
10. Make sure you updated every one of the stat-* scripts! There's six of them total:

    Code Block
    stat-general

    ,

    Code Block
    stat-initial

    ,

    Code Block
    stat-monthly

    ,

    Code Block
    stat-report-general

    ,

    Code Block
    stat-report-initial

    ,

    Code Block
    stat-report-monthly

11. You will first need to run the initialization script to gather data about all of the past months (back to the month and year specified in

    Code Block
    $start_month

    and

    Code Block
    $start_year

    ). This script only needs to be run once, though you can rerun it if you needed to (in which case it would overwrite its past results). Its output is a set of historic data files which can be used by stat-report-initial to generate historic HTML reports.
    • dspace

      Code Block
      /bin/stat-initial

      (Note: For Windows, you may need to run

      Code Block
      perl

      dspace

      Code Block
      /bin/stat-initial

      )
12. Next, generate a series of monthly reports for all of this past data by running the corresponding stat-report-* script. Again, this script only needs to be run once, though you can rerun it as needed. Its output is historic monthly HTML reports.
    • dspace

      Code Block
      /bin/stat-report-initial

      (Note: For Windows, perl dspace

      Code Block
      /bin/stat-report-initial

      )
13. The other four scripts should be run at least on a monthly basis, though you may even want to schedule them to run on a nightly basis in order to keep your statistics up-to-date at all times. These scripts will generate a report based on the activities this current month (stat-monthly and stat-report-monthly) and a general report aggregating all activities in the history of the repository (stat-general and stat-report-general). As in the examples above, the stat-report-* scripts should always be scheduled to run after their corresponding stat-* script.
14. For Linux or Mac OSX, you can schedule the scripts to run by adding a

    Code Block
    cron

    entry similar to the following to the crontab for the user who installed DSpace:

    Code Block
    30 0 * * * [dspace]/bin/stat-monthly

    Code Block
    35 0 * * * [dspace]/bin/stat-general

    Code Block
    30 1 * * * [dspace]/bin/stat-report-monthly

    Code Block
    35 1 * * * [dspace]/bin/stat-report-general

15. (The above cron entry would schedule these scripts to run nightly. stat-monthly and stat-general would be scheduled for 12:30am and 12:35am respectively. While their corresponding report scripts would run at 1:30am and 1:35am respectively. Note: You would need to change dspace to the full path of your DSpace installation directory.)
16. For Windows, you should use Windows Scheduled Tasks to schedule those same commands at an appropriate time of day. Just remember to schedule stat-monthly and stat-general to run before their corresponding report scripts (similar to the Linux/Mac OSX example above). In addition, you may need to call Perl explicitly (e.g. perl stat-monthly)
17. After running all statistics scripts, your DSpace site's statistical reports will be available immediately at

    Code Block
    http://web-address-to-my-dspace/statistics