Redefinition. I hope you have read the article by John Allen Paulos which I had referred to in the last post. A valuable lesson for us - changing the way a particular item is measured can affect the tend that the chart depicts. Maternal death, Autism are some of the examples that John presents.
I wrote about the deportation of Latinos and discussed the chart published by the Pew Research Center - not a great chart. The redeeming aspect was that a critical aspect was discussed in the report; the fact that the statistical increase could partly be the result of change in the way deportations are computed.
The chart above I believe can be improved in two ways, removing all those data labels and clearly indicating the possible reason of the statistical variance in the graph itself. Here are two alternative presentations we can consider;
or,
The second version is a much cleaner version which clearly brings out, both the overall increase in numbers as well as the increase in non-criminals deported.
Can you think of other alternatives?
The chart above I believe can be improved in two ways, removing all those data labels and clearly indicating the possible reason of the statistical variance in the graph itself. Here are two alternative presentations we can consider;
or,
The second version is a much cleaner version which clearly brings out, both the overall increase in numbers as well as the increase in non-criminals deported.
Can you think of other alternatives?
I thought a Marimekko chart would be a useful one if the aim is to compare the ratio of criminals to non-criminals, giving a little (but not significant) weight to the growth in immigration.
ReplyDeleteHi Anand
ReplyDeleteThe main message was the somewhat spurious growth. That is the reason why the total removals have been repeated in both the charts. The secondary message is the trend in the component and the proportions. The second option above fulfills this objective.