One database, one decision, two very different charts

One database, one decision, two very different charts

This post has taken a few days to percolate up through my brain. Not because it is profound, but because it is subtle.

tableau-chartsA single, subtle decision (which has no right or wrong answer), led to two drastically different results on TheDenverChannel.com and DenverPost.com.

In light of the measles outbreak, our reporters were cooperating to obtain data on immunization rates from local school districts. I was asked to join the project to build a visualization of the data for our site, just as the Post’s reporters asked a teammate to build their own.

Both of us decided to use Tableau to create those searchable databases. We did select different graphs, but that is not why our results were so different in the end.

What led us each to different results was an editorial decision on how to organize the data. The newspaper team followed Colorado law, which allows for children to be “compliant” with vaccination rules even if they aren’t vaccinated. Meanwhile, we decided to simplify the data and group the children who’ve filed for exemptions with those who are non-compliant because they haven’t received the shots or haven’t proven that they are compliant.

Neither approach is wrong, per se, but here is a quick argument on behalf of our decision: If parents are concerned about the number of kids who might catch or transmit the measles, it does not matter if unvaccinated children have filed the proper paperwork. We hoped our decision would make the chart more rapidly consumable for parents.

Here are embedded copies of both Tableau charts, starting with the one I built.