Snapshot Formulas

A data delivery system should, whenever possible, provide the end user with data in the exact format they need. Most systems, however, provide files which are simply counts of all possible variables only, such as the number of females age 35-39 in an area. The entire delivery of data consists of a large number of variables in this “raw” form, leaving the user to undertake the calculations necessary to render the data useful for their particular application.

AGS has always taken a different view and provided simple tools that manipulate the raw data so that users need not have the burden of manipulating the data themselves. For example, we could present the variable as a percentage, but the question arises as to a percentage of what? We can immediately think of three candidates –

• % of the population age 35-39 who are female
• % of the female population who are age 35-39
• % of the total population who are both female and age 35-39

The AGS solution has been to use variable lists or report templates which define the output data in exactly the form required, and to output only those variables that the user is actually interested in obtaining using functions like median() and percent(). With the release of SnapShot, the range and function of these formulas has been significantly improved. Over the coming weeks, we will be exploring some of the more esoteric formula functions in some detail. For now, the new processor includes –

• rank() which is used with Panorama to select the largest (or nth largest) item in a list, allowing an easy way of extracting just the list of the top 3 segments in an area. count() tells you how many Panorama segments can be found there.
• Mathematical functions for standard transformations including log(), exp(), ln() and sqrt().
• New measures such as diversity() and inequality()
• interpolate() which will estimate a variable for any intervening year between data points so that you can match your 2014 survey with appropriate demographics.
• class() which allows the value of any variable to be converted to a user defined classification system, such as rescaling our CrimeRisk data to a five point scale.
• Modeling support including logical test functions and a regression() function which will output regression estimates given a set of coefficients and independent variables.
• percentile() and zscore() functions which can be used for benchmarking purposes or for analytics purposes as they compare any variable with expected values from the ‘universe’ of areas of similar area nationwide.
• distanceweight() which creates our unique DATA (distance adjusted trade areas) analytics tools.

Expect more from your data provider than one size fits all datasets.