When performing data mining or data analysis, we often look internally to see why there may be a spike or anomaly in the results. Did we have a marketing promotion? Is this a historical seasonal trend in our industry? But sometimes we have to look outside the proverbial data box and look at external factors.
This article describes how one provider noticed a decrease of incoming patients to Children’s Hospital of Wisconsin when the Green Bay Packers game was on the air. After the game had concluded, however, the volume of incoming patients spiked.
Now there is not enough information in the article to come to any real conclusions. It is all just speculation. However, it does underscore the point that not all changes in data are related to what happens within an organization. Some changes are due to external factors.
Other examples include customer satisfaction ratings at theme parks which have been seen to have a close correlation with the weather. That one isn’t surprising when you think about it. And the weather is neither controllable or easily predicted so that a theme park operator could do much about it.
Gas prices, on the other hand, have a strong correlation on a variety of factors. A data miner once was mining welfare fraud data, and when he added in historical gas prices, he could see a strong correlation between that and fraud. When someone asked him “what made you even think to add gas prices in” he responded, “it is a good reflection of how the overall economy is doing at a given point in time.” Good economy = less fraud.
So as you are sifting through your own data, think about external factors and how they may be impacting your organization.