Now that the latest instalment in Computers Vs. Humans has come down to a comprehensive victory by Watson against more fleshy counterparts in US quiz Jeopardy, maybe we’re finally reaching the moment in history when we as a species have nothing left to contribute but to sit in front of TV screens, watching our new silicon-based overlords duke it out to win cash prizes?
Well, it may not be time to cast off the shackles of work, settle down on the sofa and order in the extra-large nachos yet. Technology advances and statistical methods are providing the tools to take a lot of the automation and pain over both mundane and sophisticated tasks, but computational power and current levels of artificial intelligence alone cannot match the creative capabilities capable by the human mind.
When developing Reportal, we have always looked to provide tools to explore the meaning of your data further, to allow actionable insights to be applied from the results. Last year, we added correlation analysis to the platform and in the upcoming release of Version 16 we have added the capability to perform simple and multiple regression analysis.
Linear regression is one of a number of statistical tools that are used to understand how an objective is driven by other variables. Perhaps a manager would like to find out which conditions have the greatest effect on employee satisfaction so that he/she can implement changes to improve the working environment in the company. An analyst would identify a number of factors or dimensions such as salary, relationship with their boss or level of training that are believed to contribute to the overall satisfaction of an employee, and then by conducting a survey this information can be combined and used in a multiple regression analysis to build a regression equation.
The stronger the effect a variable (e.g. salary) has on the objective (employee satisfaction), the more “important” the variable is, and the analysis will allow the user to model the impact that changes in the variables will have on the objective.
The important part is that these methods, like many others when considering statistical techniques, are simply tools for a craftsman; they can discover valuable information but it takes the art of the analyst to select the inputs wisely and weave this information with human understanding of the problem to create a compelling story of the data in context and how it can be acted upon.
So for now, technology rightfully remains our servant rather than our master, and we can make use of its increasing power to solve real issues from our customer data then coming up with the answer “Who is Bram Stoker?" to win a game show.