Primo Reporting: Using 3rd Party Software to Create Primo Reports & Analyze Primo Statistics (2012)
Download as PPTX, PDF
0 likes104 views
This document discusses using third party software like Cognos and Google Analytics to analyze usage statistics from the Primo discovery system. It describes how the University of Waterloo created PowerPlay cubes in Cognos to interactively analyze and visualize data from the Primo Reporting Schema views. Impromptu reports were also created in Cognos for specific predefined reporting needs. Google Analytics was used to supplement Primo statistics with additional web usage details. Working with Ex Libris support helped clarify questions about the Primo data views.
Primo Reporting: Using 3rd Party Software to Create Primo Reports & Analyze Primo Statistics (2012)
- 1. Primo Reporting: using 3rd party software to create Primo reports & analyze Primo statistics Alison Hitchens & Gail Sperling University of Waterloo Presented by Alison Hitchens at ELUNA 2012
- 2. Agenda Primo Reporting Cognos Software Suite Playing with the PowerPlay Cube Impromptu Reports Working with Ex Libris Support Google Analytics
- 3. Why metrics? How much is our system being used? How is our system being used? Challenge assumptions Information for testing Raise questions Justify decisions CAUTION: make sure you understand what the data is based on!
- 4. Back Office Reports There are out-of- the-box reports available in the Primo Back Office
- 5. Back Office Reports Interesting statistical reports such as the Click- through statistics Can filter by institution, user group, source, start/end dates Detailed results give you information on click type, view, count, click-value, institution, user group and on/off campus
- 8. Creating local reports The Primo Reporting Schema (RPT00) gives read-only access via Oracle Views to: PNX-related information Search-related information User-related information and events System monitoring events You need to install the RPT00 schema on your Primo server You can incorporate locally created reports into the Back Office if you use Eclipse BIRT Designer
- 9. Creating local reports At TUG we create our reports using Cognos Filter on view rather than institution Flexibility to choose filters on the fly Dynamic* reports not static reports We already have Cognos expertise We focused on two views: SEARCH_STATISTICS CLICK_EVENTS *reports that can be dynamically manipulated
- 10. Content of the views SEARCH_STATISTICS scope_name scope_type search_count average_results average_search_time source_view institution on_campus user_group CLICK_EVENTS event_date event_type click_value refine, full display, help, GetIt, add to e-shelf, add a review, previous page click_count view institution on_campus user_group
- 11. Documentation Primo Reporting Schema and User-Defined Reports (Primo Technical Guide, Chapter 14) How to install the RPT00 Schema How to install user-defined reports (if using BIRT) Description of each of the reporting views
- 13. Cognos Software Suite PowerPlay OLAP (online analytical processing) software for analysis PowerPlay Transformer The modeling tool used to define how a cube will look Impromptu Web Report A report authoring tool for detailed list reports
- 14. What is a PowerPlay Cube? the data to be used for statistical analysis by PowerPlay is built into a PowerPlay Cube PowerPlay Cube data is static but can be built or refreshed regularly as required when analyzing data the user is not affecting response time on the database
- 15. Measures Measures are the numerical data to be analyzed For the Primo cube we used: Primo Count = CLICK_EVENTS Facet Count = CLICK_EVENTS with facet info Search Count = SEARCH_STATISTICS DeepSearch kept separate because of problem
- 16. Dimensions different ways to view the numerical data when, where, who, what e.g. date, Event Type, View, Institution can be hierarchical Multiple categories can be grouped together in a folder
- 17. Slicing & Dicing Dimensions
- 18. Filtering in PowerPlay A filter narrows the focus of the report and does not have to be included in one of the rows or columns A filter can be applied by Selecting a category in the dimension line Right-clicking on a category in the dimension viewer and selecting the filter option Clicking on a hyperlink in a row/column category
- 19. Other Features of PowerPlay Once a report has been designed it can be saved and will refresh with updated numbers after the cube has been rebuilt Measures can be used in rows or columns Graphs can be easily created Reports can be saved in .pdf .xls or .csv formats
- 20. The Power of PowerPlay Statistics at users fingertips No need to ask the expert to tweak a report Web no workstation installs or updates No impact on database and Primo response time Saved reports update automatically
- 21. Creating a PowerPlay model for PRIMO Queries are run against the database views CLICK_EVENTS and SEARCH_STATISTICS Calculations are included in the queries to create the necessary dimension hierarchy Facet counts are limited to the first facet of each type if multiple facet values are applied for the same facet type only the first one will be applied
- 22. Example query against RPT00 to create the event counts Includes calculations, e.g. group into weekday vs. weekend; e.g. change University of Waterloo to WATERLOO Primo totals = Primo Count + Facet Count
- 23. Designing the cube Involves a lot of back and forth clarification between requester (me!), developer (Gail) and Ex Libris Support Cube is tested on our Cognos test server so that we can check the model and make sure we can combine dimensions in useful ways with expected results
- 24. PLAYING WITH THE POWERPLAY CUBE
- 25. Example PowerPlay reports Basic searches by year and day of week Searches by institution in fiscal year 2012
- 26. Video demo of Power Play (hopefully) we are going to switch to a video demo to show the easy and interactive nature of the PowerPlay cube The following slides duplicate many of the steps in the video
- 27. PowerPlay: OPAC Primo Default view of the cube with first two dimensions displayed.
- 28. Clicked on 2012
- 29. Clicked on Add User Data
- 30. Replaced date column with UW Views
- 31. Suppressed zeros
- 32. Nested UW user groups
- 33. Playing with facets We created a separate measure for the facet counts We created a separate dimension for each facet This allows us to: Look at which facets are used in combinations Drill down to the specific facet values that users are choosing
- 34. Facet combinations: resource & date Resource Date Created individual dimensions for each facet type to allow comparison
- 35. Facet drill-down: LC Classification & UW User Group
- 37. Impromptu list reports Staff request a report describe purpose of the report, list data elements, criteria for filters, and sort order Developer creates the report and publishes it to Cognos Staff can view the latest report or run the report from the Web whenever they need it
- 38. Sample report: Primo tags list (based on the PNX_TAGS view) View the latest or choose run to select different options This report has 2 filters: Starts with Status
- 39. Sample report: Primo tags list At the moment we can link to our OPAC to see the title; in future we will link to Primo
- 40. WORKING WITH EX LIBRIS SUPPORT
- 41. Finding Answers Shout out to Ping Gao at Ex Libris NA Support! When does the data update in RPT00? if using real-time online transaction processing then changes occur once data is written to Oracle Clarified that the click-event for basic search and advanced search is based on the searches on those tiles not the click on the link to those tiles Clarified the correct click_event code for basic search Confirmed differences between GetIt!, GetItlink1 & GetItlink2 (one is based on Primo 2; the others on Primo 3)
- 42. Finding Answers In the search_statistics view there is scope_type and scope_name, neither of which indicate blended search Blended search is included in local Why are we seeing both ONLINE RESOURCES and ONLINE_RESOURCE for the top-level facet? It was used briefly by one view. Ping tracked this down! What do X field mean???? e.g. Database, Network, Data Gained information to filter out fields that are for Ex Libris internal use only
- 43. Outstanding issues Click_event for recommendations Doesnt match the number of times we click on the recommendations tab or on a recommendation (SI 16384-307240) Click_event for DS is returning the number of results from the deep search rather than counting how many searches the users have done (SI 16384-291077) None of the views store data for the pre-search filters None of the views store data on which tab is being used It would be nice to have more definitions in the documentation
- 44. GOOGLE ANALYTICS
- 45. Why add Google Analytics? GA generates detailed statistics about visitors to a website and it is free! (www.google.com/analytics) Can supplement the information from other sources But cant be combined with other dimensions in Cognos If you make changes to your site you can profile the effects on usage Aids troubleshooting Aids website optimization You can profile corresponding time periods, for example Winter 2012 semester and Winter 2011 semester
- 46. Why add Google Analytics? Includes information such as: Visits, unique visitors, language, country, browser, operating system, mobile device Where the traffic is coming from, e.g. direct, from the librarys home page, from SFX Can do virtual page tracking to access information on sign in, my account, tab use *Requires JavaScript code on every web page that you want to profile
- 47. Browser information for Jan. 3-Apr. 27, 2012
- 48. Mobile devices Jan. 3-Apr. 27, 2012
- 49. Use of remote tab
- 51. Thank you! Questions? Gail Sperling Cognos questions gesperli@uwaterloo.ca Alison Hitchens Primo content questions ahitchen@uwaterloo.ca
Editor's Notes
- Challenge assumptions (e.g. no-one uses advanced search) Information for testing (e.g. we need to make sure things work in Chrome) Raise questions (e.g. we need to find out why few users are using the add a review option) Justify decisions (e.g. financial formulas for consortia based on use)
- -slow!
- (see Technical Guide ch. 14)
- (see Technical Guide ch. 14)
- Ch. 11 of the Back Office guide describes the existing reports
- Transformer = design of the tool! Many queries may be used to create a PowerPlay Cube
- 14
- 15
- 16
- 18
- 19
- Need to control for the variances Powerplay still part of Cognos 8-10
- Can be done on the fly or I can decide to save it for later; will be updated when the cube updates.
- Simply click and drag to replace the dimension
- Most importantly, if I like this report I can save it to my use box and run it whenever I like!
- *delay of approximately 75-90 min. *we saw Basic, BA, Basi etc.