Opening up our web analytics for dynamic, public display

Who is using Open.Michigan? What are our most popular courses? How much do people use our content? These, among others, are questions we care about. And we know that some of you are as curious about some (or all!) of our usage data as we are. The numbers have been there, in our web analytics accounts, for several years — just buried. But Open.Michigan’s usage information now begins to become, fittingly, more open, as we’ve just released the first iteration of Open.Michigan’s Dynamic Analytics Project.

Over the past couple years, we have made a more concerted effort to track and examine data regarding the use of our materials. The Dynamic Analytics Project is an effort to openly share Open.Michigan’s usage data in useful ways. What does “better” mean? What is a “good” way to share data? We are asking these questions too.

This release is only the first phase of our dynamic analytics project. In future phases, we aim to identify and share meaningful patterns, provide easy-access evidence to how many people have viewed and downloaded content from our site, and remove the barriers to this data that currently exist.

Let’s dig up the numbers

People across the University of Michigan community and around the world create openly licensed content that Open.Michigan either hosts or references to on a landing page. Often, some content creators will ask us for these numbers; say, “How many times has the course I created been viewed in the past month?” They can’t all log in to our Google or YouTube analytics account. Though individual YouTube video and channel view counts are public, it’s quite a time-sink to aggregate the views across all 22 videos in the ARCH 324 – Structures 2 course, for instance. Open.Michigan’s dashboard on Geckoboard provides aggregate data, but that’s only the tip of the iceberg. We shared some of the African Health OER Network usage data through the A2DataDive and Health OER Tech workshops, but that was also just a few puzzle pieces.

With this new feature, any visitor on our website can finally answer questions such as “How many people have taken a look at this course page?” without requiring an Open.Michigan intermediary.

A look at the features in release 1.0

You can now see the current display of some usage numbers on any course page. Hit this button: image of the button to click for analytics display
and you’ll get a display that looks something like this, a screenshot from the M2 Curriculum Cardiovascular sequence:
image of analytics display on m2 cardiovascular page

You can see the number of times the course has been viewed, the number of times course materials have been downloaded, and, if there are any YouTube videos for the course, the number of times all those have been viewed.

How it works

This display is enabled via a new module built for OERbit, our custom Drupal platform. We have been learning things about Drupal and user interface design every day. I’ve had the chance to write scripts similar to others I know well. I’ve read about things ranging from database design to JavaScript charting libraries.

We use the Google Analytics API and the YouTube Data API via custom Drupal modules (our code is available on GitHub) to gather the data that we currently show and integrate it with the existing structure of our site (more technical details to follow).

The particular numbers we currently show, listed above, were chosen as the focus for this first release because they are the answers to common questions we receive from creators of our published content. We wanted to empower content authors to answer those questions themselves. We believe that consumers, remixers, and researchers of open resources should have access to their usage stats, as well as the creator and content publisher.

We want your feedback!

This is just the first release — a fuller display of analytics and updated features, such as geographic usage information and charts to indicate change across time, is planned for future releases.

We want to know more about what you — our audience — want to know about usage information from Open.Michigan. Please share your input by completing this short survey.

Future plans

This is an ongoing learning process for us. We’re examining best practices for data sharing with respect to content and technical architecture, we’re questioning how to make usage data sharing most useful, and we’re investigating how to display things clearly without obscuring other information. I’m enjoying how the building of this project so far has been all about learning and asking interesting questions, and I have learnt a lot myself throughout the process. We’ll keep looking at what numbers are surprising to us and why, and we’ll keep you posted on our discoveries.

Our project goal is to increase ease of access to usage data from Open.Michigan (more details also to follow!). We aim to provide answers to more questions, like where is this course most commonly accessed?, using the YouTube Analytics API. And of course, better charts and visual representations will make the information gleaned from usage data more easily accessible as well.

This Dynamic Analytics Project means we’re looking to find and share meaningful patterns in data, to answer questions, and to make it possible for anyone to do their own analyses with Open.Michigan data. Easy, open access to usage numbers in a clear display is the first step, and the possibilities are exciting.