[Update] On Thursday, July 31 I presented this little webinar as part of the American Evaluation Association’s coffee break webinar series. If you missed it, AEA was kind enough to give me a copy of the recorded webinar to share with you.
Below here you’ll find most of the cartoons used in the presentation. You’ll also find interactive versions of the two samples I mentioned.
If you have any questions, please let me know in the comments.
Stop Wasting Data
We all waste a lot of data. What do I mean by wasting data? Here’s my definition…
Wasting Data = Failing to deliver potentially nourishing data in a consumable format.
-To tweet this just click this link.
Evaluators see quantitative data differently
Disaggregated data can be incredibly valuable in an evaluation context. Just think about outliers. For statisticians you need to find them because they can mess up a model. For evaluators these are programs, states, cities, or participants that could provide you with some valuable information.
The canyon of uncertainty
This is the gap that separates qualitative and quantitative data. It can be bridged.
Not about fancy charts
The real value in interactive data visualization is not in making fancy charts. It’s in creating a tool that communicates something you could not communicate in other ways.
Nobody wants 500 page reports
So in traditional formats, we pick and choose what we report then data dump the rest. Going interactive is the alternative, and it’s far more practical than you would think.
So how do we approach?
Shneiderman’s mantra is a good place to start.
A good overview
Is a start, not a finish. You want something your audience can dive into and find what they really need. Scatterplots, maps, dot plots, and line graphs make really nice starts.
Disaggregate your descriptive
What are the important numbers behind the numbers that you already report?
This is a dashboard I created using fake data based off of a real measure. It shows an indicator (average star rating) broken down into pieces (Facility, Star Rating, Enrollment). There is additional information that could help so that was added as a filter.
Breaking down the data allows you to get the answers to different questions. What specific child care facilities are not doing well? The indicator looks great for one of the counties, but where could they improve?
If you have trouble seeing this on my site, you can see it on Tableau by clicking here.
To Filter or to Color
What goes where depends on your intent and your audience.
Color what you would like to compare directly within the visual. Use filters on all the rest.
This visual (using 2011 – 2012 IDEA Part B Child Count Suppressed Data) shows the number of children identified as having a disability for every state broken down by race/ethnicity and filtered on age group/disability type.
The filterable excel sheet this is based on has 13,217 rows of data.
If you have trouble seeing this on my site, you can see it on Tableau by clicking here.
Are you doing interactive work? Have a question you hope I address on Thursday?
Let me know in the comments.
Questions Anyone? Waiting for Thursday?
Anyone doing any interesting dashboard work?
Is the webinar recorded? I can’t make it, but would love to watch it later!
I believe that AEA records all of their coffee breaks and posts them on their site. Not sure how quick it will go up, but it should go up eventually. One note: the archives are open to AEA members only.
http://comm.eval.org/coffee_break_webinars/CoffeeBreak/
We recently started incorporating simple interactivity into our annual evaluation report, which includes to-date findings from an ongoing initiative-level obesity prevention evaluation. We found that simple pop-up boxes are an effective way to add details without distracting from the report’s main conclusions or adding a lot of extra length to the report. This kind of interactivity can be easily added to PDFs and widely shared with stakeholders. Check it out: http://cphss.wustl.edu/Products/Documents/HAC_9_2014_Evaluation%20Report_FINAL.pdf
For optimal viewing, download the report to your computer and open with Adobe Reader or Acrobat.
Thanks for sharing Rachel, it’s a sleek looking report 🙂
I think it’s interesting how much web design is sneaking its way into traditional reporting.
Hi Chris! Thanks for the Coffee Break! You mentioned a few other software names to do this type of work, besides Tableu but I wasn’t able to catch them all. Do you have a list anywhere you could share? Thanks again for your work!
In no particular order there is…
http://www.qlik.com/
http://spotfire.tibco.com/
http://www.microstrategy.com/
That’s if you want an off-the-shelf visual analytics tool. If you can code there is more out there but only so many evaluators I know can code.
Great presentation Chris! I always love your use of cartoons. I really appreciate your tips for not wasting data. I’ve used Tableau in the past with clinical data (I presented at their conference last year which was SO fun!) and am currently trying to get the school district I work for now on board for interactive data. I shared your tips with my team and hope to put them in action soon! Thanks for sharing!
Thanks Jordan 🙂
It can be pretty tough to bring people on board. Let me know how it goes.
Best way I’ve found so far to “sell” interactive is to create very specific demos. If the data is not public, and you don’t have a Tableau desktop license, use dummy data. I also think the more live examples of the benefits the better.
Regret that I couldn’t attend during the original coffee break. However, thank you VERY much for putting this up for the public (thanks, AEA!). I love that all of your work is easily digestible and is so organized so that people from various areas and skill levels of evaluation and research can get on board and learn from you!
Very kind words Sara, thank you 🙂
Thanks for this informative session, Chris. Continuing with the theme of making data nourishing and consumable… your cartoons make it palatable, too!
Thanks Patty 🙂