In 2004 Hurricane Frances was approaching to hit Florida’s Atlantic coast. But far away from Florida in Bentonville, Arkansas, executives at Walmart stores realized the opportunity and decided to use predictive technologies in order to assist to stock up the stores before Hurricane and make profit.
The Chief Information Officer (CIO) of then Walmart pressed her staff to forecast based on the sales of past Hurricane which hit few weeks earlier. CIO felt that equipped with Terabytes of data on shopper history which was stored in Walmart Data-warehouse, CIO realized it’s time to predict rather than take reactive measure after occurrence of the event.
Usually obvious results such as an increased purchase of bottled water and torch lights are obtained. CIO was intending for somewhat useful insights.
The team of analysts helped the Walmart to take data driven decision by analyzing the data and retrieving patterns from past hurricane situations. The experts mined huge amounts of data at their disposal to realize that Pop Tarts has increased sales during Hurricanes in fact 7 times their usual sale rate and they also realized Beer was top selling item during pre-hurricane duration. This helped Walmart team to stock up to cater unusual local demand of products and make profit.
In my opinion there is always a hidden story in the data. An effective way to extract the story is to understand the problem and analyze it according to the situation. One size doesn’t fit all. Every data problem needs a different perspective and a world view to solve it.
Book: Data Science for Business by Foster Provost and Tom Fawcett