Measure Factory Overview

Measure Factory is an add-on to the Diver Platform that automates the application of domain-specific rules.

NOTE: Measure Factory implementations require a special license. Assisted Analytics (AA) within Measure Factory also requires additional DiveLine licensing.

The diagram below depicts the major elements of a Measure Factory implementation.

Measure Factory Overview 71

Your organization can implement Measure Factory by purchasing a DI application, such as Hospital Operations or Productivity. The deployment is done by DI consultants, while you validate the data. Standard dashboards and reports are part of a hospital Measure Factory implementation. You can also implement a Measure Factory from the ground up. Either way, the topics in Workbench Help are intended to help your organization define, use, and maintain measures. Information on building dashboards can be found in the separate DivePort Administrator Help.

What is Measure Factory?

The Measure Factory is an automated rules engine. It applies all the rules associated with input data sets and provides transformed output data sets. The output data sets allow you to present this transformed data to end users. One important feature of the Measure Factory is the automation of the rules application, which allows the developer to concentrate on rule development and not worry about the rule processing order.

What is a Rule?

A rule is the application of a particular business rule’s logic that transforms one of the input data sets. The business rules are based on the organization’s data governance and ensures that all end users work with a consistent view of the data. There are six rule types. They can be as simple as counting the records of a column from an input cBase (source rule) to calling a complex external process (plugin rule). There are rule types for calculating columns (calc rule), bringing in columns from text data sources (lookup rule) and applying Boolean flags (flag rule), as well as one that allows you to bring in data from another input cBase (link rule).

What is a Measure?

A measure is a calculation that is derived by applying business rules to your source data. Each measure applies one or more rules, which may span multiple data sets. Typically the result is a number in the form of a ratio or percentage. It could also be a date, Boolean, or string. A measure's definition includes written information that provides context about what the measure represents and the logic used to generate it. Measures are associated with a time period so that you can compare changes over time.

Understanding Rules and Measures

A rule is like a column in a data set. Each record has one value for each rule.

A measure is a way of summarizing or aggregating rules for a given set of records.

Applying one rule to five records gives five values. Processing five records through one measure gives one value, which can then be split over a dimension—for example, by diving on time or facility.

What is a Scope?

A summary scope is a limit applied to the source data set. The scope defines which data set a measure is calculated from and what, if any, limits apply. A measure does not require a scope; measures without a scope can infer one from the measures used to define it. A measure with a scope can reference rules in the corresponding data set, and other measures using the same scope. A measure without a scope can reference other measures, whether or not they have a scope.

NOTE: Most measures have a scope explicitly defined as part of the definition.

One data set can have multiple scopes associated with it. The data set is read once and these scopes are an abstraction layer into portions of the data. For example, an Accounts data set can have an Admissions scope and a Discharges scope. Both of these scopes are a count of accounts and return different numbers because they are limited by each scope’s filter.

For a more in depth overview, see this white paper: Measure Factory Overview.pdf