Analytics as a discipline is certainly a hot topic. In reading the book “Weapons of Math Destruction”, I discovered that analytics and its applications surround us every day and we don’t even know or acknowledge them.
What is the right price to pay for a house or a used car? Who should be admitted into a college program and who should be given the home loan and at what interest rate? What makes a good customer for a bank or a credit card company?
Implicit in these questions is a bit of a moral dilemma. The type of moral dilemma brilliantly discussed in Cathy O’Neil’s book on Weapons of Math Destruction. Here is a simple example: what kinds of data should be used in calculating a credit worthiness score? Should your zip code be part of the formula? Should your shopping patterns be included? Should people who eat lunch at Arby’s be punished on their FICO score because someone somewhere found a correlation between eating at Arby’s and defaulting on a loan?
The question on zip code is a bit of a larger moral dilemma. Zip codes can certainly be sorted by median income and other measures of “wealth”. Is it fair to discriminate against people on the basis of residential zip code? Should people be grouped into these mathematical “clans” and is it fair to generalize for entire populations in this way?
As you design your algorithms, ask yourself: How transparent is the process? How consistent and fair is the process? Are the input variables relevant and are you measuring actual measured behaviors or something tangential? I recommend the book to those who are in any way working in the analytics space. It is alarming and will make you think. Morality can go beyond cheating on a test or fudging your income taxes, it can also apply to how we treat each other in the world of analytics.