Hard Science, Crystal Balls and Educated Guesses - The Art of Predictive Analysis

The need to predict the future has probably existed since the dawn of time.The reading of tea leaves in ancient China, astrologers advising Kings, and even your Mother patting your hand when you were young and crying, telling you it would be all right.

Just after the dawn of time, probably about 10 O'clock, models were created to fill the need and desire to know the future about to what was going to happen in business, love and war and all the other various trials and tribulations of the human experience.

Models are not always complex or even mystic.A simple retail model example for evaluating a trade area might ask these simple questions.Does anybody live there? Can you get there? Is there any money in the area? If the trade area model gets a "no" to any of these, the model says "next".

Perhaps the most classic simple model came from comedian George Carlin's "the Hippy Dippy Weatherman" when he said "the forecast for tonight is dark, followed by scattered light".A model easily validated.

Dice

If you would like to see a dictionary of predictive terms and methods go to http://www.homestead.com/ssdivining/MethodsDictionary.html. One of my favorites from there is Cleromancy - predicting the future using dice, or Crystallomancy - using a crystal ball, which in terms of confidence and accuracy, are no doubt right up there with "Gut Feel".

Today, a number of retail modeling techniques run from simple analogues, to network optimization, heuristic models, econometric models and there are even the scenario models used in the military intelligence business.These all move the ability to assess the future much further ahead than the techniques of Cleromancy or Crystallomancy.

Most current modeling methodologies rely on visualization to portray the model's results.Examples might be; Business Intelligence software (definition) that can let you visualize the present, past and the future based on what is currently known or happening, on a dashboard.GIS software (definition) can let you see the future on a map.Competitive Intelligence software (also see www.scip.org) can let you see the relationships of data, then, now and as a possible future.Scenario modeling is predicting the future based on events, people (as players in the model) and existing environmental conditions that affect choices (for an example of scenario modeling techniques, check out this paper*).

As the theme for this month's issue of Location Intelligence Magazine is predictive analytics, you will find several different approaches to the subject.We have an interesting collection of modeling examples from retail, from BI, science and a continuance of modeling from a geo-temporal perspective. All of these methodologies stem from hard science and you will not find a "black box", any mysticism or Crystallomancy.

But in spite of the science, modeling the future represents one of the more fascinating fields of human endeavor.Prediction technology has moved past Madam Cleo, the fortuneteller, and has achieved a level of credibility and acceptance that has made it a "must have" in 21st century business.

Explore the contents and enjoy this issue of Location Intelligence Magazine.