Monday, March 3, 2014

Choosing the right visualization

As of the start of this year I became the director of sponsorships and memberships for local chapter of User Experience Professionals Association (UXPA). The first thing I envisioned that I need to learned about was what the current and past members liked and disliked about our organization. It was time to create survey to get some input. With hundreds of members I was expecting a potential problem of gathering too much data, but that is a good problem. I thought of using SurveyMonkey or WuFu Forms. Then someone mentioned Google Forms, so I found them very easy to setup and use. I created two forms, a simple one for past members comprised of three questions. I wanted to give open ended questions to let them freely share their thoughts. I created a second form for active members where I wanted to gather more detailed information. I shared the two forms with the other executive members to get their feedback. That one step was the best decision I made in the process. A couple of them had great ideas which made the forms even better. One of the board suggested I look at a government site on building surveys, which I surprising found very useful. The companion usability.gov site was even more interesting as it looks well designed and also has useful information, which was a double shocked. It is interesting how perceived notions sometimes don't help us at all.

After gathering just five days of survey data, I only have 19 responses from active members, which is disappointing but still better than guessing or having no data. I wanted to explore creating a visualization of the data collected from the survey as I used a rank order to try to figure out what members like that we are cur renting doing. This is my survey data:


Rank
Book Club
Webiners
Workshops
Community Events
Social Events
member 1 5 2 1 3 4
member 2 5 2 1 3 4
member 3 5 2 1 4 3
member 4 5 1 3 4 2
member 5 4 1 3 2 4
member 6 3 2 1 1 2
member 7 2 3 1 3 4
member 8 5 4 2 1 3
member 9 5 4 1 2 3
member 10 4 1 2 3 5
member 11 3 2 1 4

member 12 1 1 1 1 1
member 13 5 3 1 2 4
member 14 5 2 1 4 3
member 15 3 2 1 3 5
member 16 4 3 1 2 5
member 17 4 3 1 2 2
member 18 3 2 1 5 4
member 19 3 1 1 3 2

So I want to creating a visualization of this data to know what is important to our members. I used a "rank order visualization" google search to see what I could find. I found a gem in one of the UX stack exchange entries, where a discussion of using column charts, pie charts, bubble charts and even Excel spark lines was suggested. I looked at another Stats stack exchange article which was a deeper explanation on using spark lines. Then I got a bit crazy and wanted to look up existing research on how to visualization my data as it just seemed like a old problem. I found this interesting IEEE article from GaTech on using a heat map. While exploring, I then wondered how people visualization Likert scales, which I did not use for my surveys but it is definitely related and maybe be helpful. I found this great article which I need to file and use in the future. I especially like this one:


I am also quite fond of this paper on correlating rankings and then realized I was getting a bit off track. I went to the site for my favorite graphing package called HighCharts to see how easy some of these could be achieved. I really like bubble charts but after looking at the data I realized that this is not the correct visualization to help me. Finally after a couple of minor diversions, I went to Apple Numbers to add the data and created a summary table of the counts of each ranking by type and came up with this simple column chart that tells me every I need to know:


I just look at the column chart and find the largest bars for each type and I am done:

1st - Workshops
2nd - Webinars
3rd - Community Events
4th - Social Events
5th - Book Club 



No comments: