Setting Defaults in Tableau

Here’s a quick tip/trick to using Tableau that I don’t see mentioned very often: Setting default properties for dimensions and measures.

Recently I was sitting in through some Tableau training at work, and the trainer was talking about setting up custom sort orders by rearranging the levels of a dimension in a table. This is what I typically see people do, and I admit that I usually do it that way myself. But then he pointed out the ability to set a custom order as part of the default properties for the measure so you don’t have to reorder things every time you use it. I vaguely remember seeing this mentioned somewhere before, but I can’t remember where and don’t think it’s mentioned often enough.

So, how do you do this? Well, in a worksheet, right click on the pill for the dimension or measure. Towards the bottom of the menu there should be an option for “Default Properties”. Now, the exact options you can set are going to depend on whether you are working with a dimension or a measure.

For Dimensions, you can set defaults for “Comment”, “Color”, “Shape” and “Sort”. In the realm of higher ed data, the sort can be useful for setting the default order in which Faculty Ranks or Student Classification or Level should appear. These are two examples where the default of alphabetical order is usually not the order you want, so a custom sort is needed. You can also set custom colors here to make sure they remain consistent across multiple sheets and dashboard you might create using that data source.

For Measures, you can set defaults for “Comment”, “Color”, “Number Format”, “Aggregation”, and “Total Using”. For measures, the default Aggregation is something to definitely keep in mind. Tableau likes to default things to SUM(), but there are situations where you might prefer to always show an average or count instead. Likewise, being able to set things to default as a percentage (under number format) can be useful, for example with retention and graduation rates in accountability data.

Again, I admit to not using this option very often, but it’s something that can definitely save you some time and possible frustration later down the road.

Tip within a tip: One thing that I did bite my tongue on during the training was a comment about the color of the “pills” in Tableau. In general, dimensions are blue and measures are green. However, the color is actually for distinguishing between discrete (blue) and continuous (green) values within either a dimension or a measure. This distinction determines how Tableau treats the variable when you move it onto a sheet. I’ll probably post more about this (with a link to the helpful Tableau help page) since sometimes getting Tableau to do what you want requires converting from continuous to discrete (or perhaps the other way around).

THECB Data in a (more) Usable Format

If you’ve read through some of my previous posts, you know that I tend to use Accountability data from THECB for several of the things I try in Tableau. One challenge has always been the way data is formatted by the Accountability system. It outputs data with one row per institution and separate columns for each measure and year. And of course, sub-levels of each variable are part of the measure names. Needless to say this format doesn’t work very well when trying to use the data in Tableau, or pretty much any program.

My early way of dealing with this was to simply pivot the columns into rows. This created a somewhat usable version of the data with fields for institution, year, name of the measure, and the value. Of course this came with it’s own set of problems. It was possible to compare various institutions on the same measure, but it never worked as well as I would like. Also, it was easy to accidentally include values from multiple measures in an analysis because all values were stored in the same field and you had to use filters to narrow things down to the specific values you were after.

Over Spring Break (week before last), I decided to sit down with my Excel macro and try a new approach to reformatting the data. Long story short, I was able to create a format for the Accountability data that, so far, has been significantly more user-friendly. I’ll be making use of this reformatted data in some of the Tableau examples I’ll be posting over the next couple of weeks. For those that might find it useful here’s the Excel file:

Reformatted Accountability Data (~45 MB)

Now, I should mention that this data is just for public universities and not all institutions in the state. I haven’t spent a lot of time verifying all of the data, so if you need “official” data, you should probably use the Accountability System, or at the very least double check the values you get from this spreadsheet against it. The first few columns include the name of the institution, FICE code, period/year, category, and sub-category. After that are the primary measures from each area of the Accountability system. If time permits, I might get around to providing better documentation for it, but hopefully it’ll make sense to you if you’ve used the Accountability system before.

US lags behind other countries in education (nothing new here)

An interesting read about the state of higher education in America. In a nut shell, having a degree in higher ed does give you more skills (ie higher reading and numerical literacy), but as a country we still fall behind other countries. Honestly, the finding that the US is behind other countries in terms of education is nothing new. The reason this study is getting attention is that it actually shows that education does actually teach people some useful skills. Sadly, up until now, this has largely been a belief without much empirical evidence behind it, especially on a global scale. So, although it is something that most people will look at and go “no surprise there”, it is useful to have the data to back up the claim.

What’s not mentioned in the article, but included in the report itself, is that we also fall behind in those skills at the High School level. To me this raises an interesting question given the debate over providing “free” college to everyone at either the 2 or 4 year level. My thought is that rather than everyone needing the college degree what we really need is to be doing a better job at teaching these skills in the K-12 arena. Let me be clear that I do not fault teachers for this problem. There are many issues that go into the quality of a K-12 education that are to a large degree beyond the control of teachers, and unfortunately that control often ends up in the hands of politicians. Having national standards is an attempt to raise the standard across all schools, but unfortunately those standards are often seen as a goal to reach rather than a minimum to maintain. If things were really working as they should in education the fear over whether a school is able to reach that minimum level of skills on some arbitrary standardized test would be a non-issue. Reaching the minimum should be a given. If it were, we could then focus on raising the bar for our students. But again, there are several other factors that come into play, so I’ll leave that discussion for another post.

Americans with bachelor degrees lag behind other nations in labor skills
http://www.pbs.org/newshour/rundown/americans-with-bachelor-degrees-lag-behind-other-nations-in-labor-skills/

Showing Change with a Scatterplot in Tableau

Recently the VizWiz blog had a post showing how to create a 45-degree reference/trend line in a scatterplot. The original purpose was comparing countries based on literacy rates for females and males. Those above the line were countries where females had higher literacy rates and below were countries where males had the higher literacy rate. From there the post went on to talk about change in a couple of other datasets. But, there was an issue in producing the correct color coding to represent the change from year to year in those other datasets. This got me thinking that I wanted to see if I could figure out why the color coding of change wasn’t working as expected in those examples. (NOTE: Description of how to get the 45-degree line and fix the issue he encountered are at the end of this post.) Of course, trying to tie things to higher education, I decided to use my standby of accountability data from the Texas Higher Education Coordinating Board (THECB) which I’ve used previously.
Continue reading Showing Change with a Scatterplot in Tableau

Likelihood of graduating as a criteria for admittance

Although I haven’t seen it phrased that way before, colleges do admit students they feel will be the most successsful. The better they are at making that prediction, the higher their retention and graduation rates will be. The question then is whether you want to use retention and graduation rates as criteria for judging a school since it encourages schools to be more selective rather than more inclusive. What would be nice is a measure of how successful a school is with students they expect to fail. This is why sub-group analysis is important.

Graduation Rates by Selectivity: Freshmen, 2007 http://highereddatastories.blogspot.com/2016/02/graduation-rates-by-selectivity.html

(And yes, this is a post from the blog I posted about a few weeks ago.)

Using a Custom Map for Enrollment

The following is a quick demonstration of a way that a custom map can be used to display course enrollment by building for a selected day and time on a college campus. Because the campus in question is rather wide, it created a situation where the visualization had to be wider than I would normally recommend. The most difficult part of this process is actually creating a table that contains the coordinates of each of the buildings on campus. The visualizaton appears after the break…

Continue reading Using a Custom Map for Enrollment