Access Problem Solving (digprob)
Usage Statistics for Resources in Vera
Austin Kim in the Libraries Systems Office has been working with the Emetrics group to design a tool that gathers and reports usage statistics for all the ejournals and databases in Vera. While not perfect, this tool is now ready for use and is accessible to anyone who is included in the all-lib email list (as are all Libraries staff).
THANK YOU TO AUSTIN KIM FOR ALL HIS TIME AND EFFORT IN CREATING THIS TOOL!
To try out the statistics reporting tool, use INTERNET EXPLORER or MOZILLA FIREFOX and go to: https://aeryn.mit.edu:8443/emetrics/report.php
Questions and Answers about the Statistics ToolDo I need to use a specific browser?
Yes, please use Internet Explorer. To save on development time, Austin focused on making this tool work in one browser only for now. [Note: some staff have reported that it also appears to work fine on Mozilla Firefox.]
What time period of data is included?
There is no data prior to July, 2004 so even if you select a longer time period, you are only seeing data beginning with fiscal year '05.
What is counted?
Only hits on Vera records or "get" URLs are counted. All of the URLs in Vera now have a prefix that looks like this: "http://aeryn.mit.edu/emetrics/count.php?" . When a Vera link is clicked on, this prefix causes the click to be counted and the count stored in a database Austin manages. If someone uses the native URL or accesses an article through an SFX button, these clicks will not be counted in the usage data. (We can gather usage via SFX through SFX, however, and hope at some point to integrate this data.)
If a user starts off by clicking on one journal from a package in Vera, and then, once at the information provider's site, uses the search and browse features of that site to move to another journal, that use will not be counted.
How can I look at an entire ejournal package as a whole, and see the titles within it, too?
The "show ejournals in packages" option will group titles that belong to a package together and will show you the total use for all the titles in the package. If you click on the package name in the browse list you will see all the titles in the package, with statistics for each title in the package.
Some aggregators like Factiva do not allow us to count at the title level because we can't construct unique title-level URLs, so these packages are counted only as "Databases" and as a collective group of hits under the Ejournals listing. The individual titles are not listed. (If an aggregation like ABI/Inform is not included in Vera as an "Ejournal" it will appear only in the databases listing.)
Please note that any title that is included in Vera as both an Ejournal and a Database, which most of our journal collections are, will appear BOTH as an ejournal and as a database in the statistics listing. The total that appears under "databases" shows how many times the URL for the Vera record for the collection (e.g. ACM Digital Library) that is coded as a database was used; the total that appears under "Ejournal Packages" shows how many times the URLs for the Vera records for any of the ejournals that are part of the package (e.g. the ACM Digital Library) were used, and how many times the collection record that is coded as an ejournal was used.
I am having trouble finding a particular version of a title that I know we have, and distinguishing one version of it from another. What is going on?
If you want to look at the usage numbers for a particular title, it is better to use the drop-down option of "search resource name" rather than to browse. You may not find all versions of the title if you browse.
If you do search by resource name, and find several listings of the same title (e.g. for International Organization), you need to click on each of the titles and move to the usage summary screen to be able to tell which package a title is in, or if it is not in a package.
Titles may appear twice, even if we do not have two versions. The reasons for this is that if the URL changes, a new record is created in the usage statistics counting database.
I see that Lincoln Laboratory use is listed in a separate column; what is being counted there?
The data for hits from Lincoln Lab is inaccurate and too low for the period of July 2004 through January 2005. The totals during that period reflect only hits Lincoln staff made during that time from within our Vera tool; since Lincoln has their own version of Vera, direct use of our Vera is not a common occurrence there. As of approximately January, however, the Lincoln counts are correct, as Lincoln has changed their version of Vera so that it includes our counting prefixes.
Why do we need this tool, when we can get usage data from most of our information providers?
We need comparable data available across all information providers; inconsistencies in how data is collected and reported by various providers means that we can't make good direct usage comparisons between products with the data offered by information providers.
Also, while most information providers do offer usage data, some don't, and we need a way to assess usage of all our digital resources.
Collecting data from our information providers is extremely labor-intensive; this method, now that it is automated, gives us needed data without having to allocate weeks and weeks of staff time to gather and aggregate data.
There are other benefits, too: this tool allows us to compare on and off campus use, and use at Lincoln Laboratory versus the rest of the campus. These are important views of our data that we can't get from most of our information providers.
It's important to note that even with all these benefits of our own tool, we will still need to rely on usage data from our information providers to understand the complete picture of usage (especially for titles within packages), and to understand how particular products are used (e.g. whether HTML or PDF formats are preferred, whether users browse or search, etc.).
I have noticed some errors and/or imperfections in the way this tool works. Is this tool still under development?
NERD and CMG tested this tool and felt it was good enough to be used by the rest of the staff in its current condition. It is not perfect, but we thought it would make sense to have some broad-based use before the Libraries set priorities for what, if any, changes should be made to it in the future. We welcome feedback.
Whom can I talk to if I have questions or comments about the statistics tool?
Contact Ellen Duranceau (x3 7562/ efinnie@mit.edu) with questions. She will try to answer them, or refer them on to others as needed. The emetrics group that worked on this tool included:
Carl Jones
Austin Kim
Ellen Duranceau
Carol Fleishauer
Marlene Manoff
Marilyn McSweeney