Analytics Evidence AGSA15 - Set data not consistent with analytic data

When you run set reports on physical item counts or portfolio numbers and do a comparable analysis within Analytics, often the numbers vary by 10% or more.

There are certain analytic queries like physical item counts that should result in the same data no matter if it is run in Analytics or as a set.

This causes reporting issues for institutions which provide annual statistics, such as ACRL.

status: open
priority: high
subject area: general

Example:

Updated post-ELUNA 2018 based on suggestions from Yoel on what fields to add to the Alma Set and fields to add to the Analytics Query.

The resulting number is still off by tens of thousands.

Analysis with MMS ID Count Distinct:

Results:

Results when MMS ID distinct count added:

Set Query:

Set Results:

Old Example:

Base volume count (analytics query no MMSID limiter)

Query example

Query example with mmsid limiter (Count(Distinct("Bibliographic Details"."MMS Id")))

Set Example:

Physical item basic count shows 566,616 books (with MMS Id limiter) 1,217,536 (No limiter) in comparison with 543,560 within sets.

Set Results

If Physical titles facet is used within Item set creation, the count shows further differences with item counts in comparison with analytic queries. Counts also show unrelated items such as ebooks in the query. Some are cataloging errors most are not.

From the IGeLU-ELUNA Analytics Working Group, Analytics Evidence Document

Please send questions or comments to the Analytics Working Group Listserv