2 Replies Latest reply on Aug 30, 2018 8:59 AM by Sarah Delaney

    Policies & Procedures: Publishing Dashboards & Data Sources

    Yosief Yihunie

      Dear Higher Ed Tableau Community,


      We recently started building dashboards in Tableau and we are trying to write a campus guide for managing the process. If your institution has any policies (procedure, criteria, guideline, protocol …) regarding publishing of tableau dashboards and data sources that you can share we would love to see them.


      Instructional program review, student leaning outcomes assessment, and college fact book are the areas we are able to cover so far in Tableau. We would like to learn from your experience in which other areas we can expand the use of tableau to meet the analytic needs of the institution.


      We are also interested to hear any guidance or policies you employ to satisfy FERPA and other similar regulations regarding the dissemination of student information.


      Thank you


      Yosief Yihunie

      Santa Monica College

        • 1. Re: Policies & Procedures: Publishing Dashboards & Data Sources
          Nicole Kraft

          Hi Yosief,

          In addition to factbook, student learning outcomes assessment, and program review, we also use Tableau for the following:

          • Faculty position request data
          • Strategic accountability matrixes
          • KPI/Strategic planning
          • Survey results (NSSE, alumni, etc.)
          • Enrollment Management (YTY enrollment comparisons, admissions dashboards, credit contribution, high school graduates data, major changers, etc.)
          • Equity, Diversity, and Inclusion initiatives
          • Non-academic department data (library, honors, recreation, etc.)

          I'm the only Tableau developer on campus so I have my own rules, but nothing written down. I follow the university brand manual using the recommended colors and fonts. Other than that, I've found a format that I prefer and stick with that when creating new projects. I'm sure universities with larger implementations may have a best practices guide to share.

          • 2. Re: Policies & Procedures: Publishing Dashboards & Data Sources
            Sarah Delaney

            Think about security early on. Some people may need to see individual student level data, but most won't. Think ahead about how to structure your site so you have projects that lock down access to full data and ability to download workbook, save changes, etc, so you don't have to think about it every time you publish. Also, if you allow people to place comments on a dashboard you don't get a notice to look at them.  (I just learned about this yesterday from someone who knows MUCH more about Tableau than me, but I haven't done any investigation myself. Comments are stored somewhere, so there should be a way to set up a dashboard of them)


            There is a virtual server admin group in the forums here and their old webinars, slide decks and sample workbooks are available.  Even if you aren't the one managing all the security, it helps to know enough to communicate with the people who are managing it. For sure remember this, you can't use row level security on a dashboard you allow to be edited or downloaded because a user can take out the filter. (It looks like you can put a data source filter in a workbook and publish the data source and reconnect to that data source, but I'm still testing)


            Every campus I've been on had people who wanted to double check my numbers.  I'd plan on being very transparent with a glossary aimed at lay users and a data dictionary directed  a people who have access to the same data. I'm doing poor man's master data management with Excel. The spreadsheet has the field name, presentation name, lay definition, technical, data nerd definition and a YN column by dashboard. I have a template documentation dashboard and copy the glossary  and data nerd dashboards into every dashboard I make and just change the filter on to use that dashboard's YN column.


            You will get sick of the color blind pallet, but if you deviate, be sure to check with a Chrome plug in like NoCoffee. I used to think that as long as I didn't have red and green in the same viz it was ok, but red /green are perceived in ways that make them look like completely different colors and those other colors can be a problem.  8% of men have color vision defects and I've run across 2 men who hadn't considered themselves to be colorblind, but commented that a viz was easier to see with the color blind pallet.


            Just in general, view this presentation from TC 17 - it has a bunch of things in it that I wish I had done to help socialize my dashboards and get feedback from the users before plowing ahead:

            ExxonMobil: So you’ve built a great dashboard, now what?


            Good Luck!

            1 of 1 people found this helpful