February 12, 2016
If the Data Isn't Right, It Doesn't Matter How Pretty Your Visual Is!
The last thing you want to do is to design an excellent visualization that engages the audience so much that it drives action. Wait...isn't that the goal? It is, unless the data is inaccurate.
Let me provide a personal experience. Yesterday I went to HomeDepot.com to look for 16 very specific "shallow utility hinges" for a wood project I'm working on. Luckily they had hinges in-stock (available for in-store pickup only)...Sadly for me, I needed 16 and my local store had only 10. So in order to get the 16 I needed, I'd need to go to 2 locations. I drove to the next town over on the way home from work, got 8 hinges there and then proceeded on to the store down the road from my house. I was quite irritated to find they actually had about 30 in stock, not the 10 that their pretty little inventory by store summary was showing. So I had gone to two stores based on their data and only needed to go to one.
In the end, the data they displayed caused me to expend more time and travel to get my hinges then if they didn't have the visual at all. I would have just started with my local store, but I trusted their dashboard. Will I trust this site feature again? No, I will Not.
As a developer myself, this was a great reminder of why data quality is paramount.
Our goal should be to drive actions/decisions with our visualizations. But we need to be certain we're driving the right actions/decisions.
For me, I lost 30-45 minutes of day. Not a big deal. However, my stakeholders at work rely on my data to make much more important decisions...and the resulting actions can have far greater consequences. Would you want a decision that affects your employment status to be made off data that may not be accurate?
Check your data, then check it again. After that, check it a few more times.