Analysis Datatypes

All database management systems and programming languages have their own set of datatypes (or data types, if you prefer). But, until now, there was no set of datatypes for use in business analysis.

I've utilized these analysis datatypes for ten years in consulting projects. I've found that people understand them with little explanation and they help make the data requirements gathering, analysis, and documentation processes easier and clearer for everyone involved, as well as providing greater consistency for data element names, definitions, and the specification of their valid values.

We're making this list publicly available to support the business analysis community. We encourage you to use it for non-commercial purposes and tell others. Please contact us with any proposed additions. We'll credit you on this page for accepted additions.

Analysis Datatypes: what, why, who, when, where, and how

What are they? An easy and simple way to categorize data elements during the analysis process. They are definitely not implementation datatypes, although they can be mapped to a particular database management system's data types. See below for details of these mappings.

Why should I bother? Utilizing analysis datatypes will help your subject matter experts to be clearer about their data requirements. By classifying data elements into these analysis datatypes, you'll improve the precision of your analysis work products, thus helping to reduce the likelihood of misunderstandings. Utilizing analysis datatypes will assist your software and database designers. In each project in which I've used them, they saved many hours of time, as well as provided more consistency and higher quality definitions!

Who? Who assigns the analysis datatypes? Any knowledgeable person, certainly the business analyst would be responsible for this. I've found that subject matter experts frequently understand them easily enough they can assign analysis datatypes themselves! Who uses the analysis datatypes? The people that design the software and database: software engineers, database adminstrators, etc.

When do we assign them? As early as practical during the process of gathering and documenting your data requirements.

Where do we document the analysis datatypes? Wherever you document your data requirements. I recommend using data requirements lists and then an analysis data model.

How do I determine the most appropriate analysis datatype? Begin at the top of the list, work your way down, and stop at the first applicable analysis datatype.

Analysis Datatypes

Datatype Name	Description	Example Value(s)	Sample attribute using the datatype
indicator	a two valued code set representing the values true/false or yes/no; also known as boolean	yes; no	hazardous material indicator
code	a constrained value set, typically used for categorizing something	M (for male); F (for female)	gender code
datetime	a point in time	11/9/2008 9:09 p.m. EST	accident date and time
date	a specific date in time	January 28, 2009	birth date
time	a specific time of day	5:30 p.m. EDT	appointment time
money	a monetary amount, including the unit of currency	4.35 US dollars	bank account balance
amount	a measured or observed amount, including the units; you may want to subcategorize amounts into distance, time, weight, etc., as was done for monetary amounts	2.1 meters	height
quantity	a count of individual items such as people, places, and things	47	attendee count
sequence number	a specific position within an an ordinal ranking (e.g., by time, by importance)	2	birth order
percentage	the ratio between two amounts having the same units or between two quantities of the same type of items, expressed as a percentage	75%	current year cost increase
factor	a number that an amount or quantity is multiplied by to produce another amount or quantity	2.3	currency conversion factor
address	a street or post office box address	157 Raymond St, Cambridge, MA 02140, USA	work address
phone number	a telephone number; includes phone, fax, and pager numbers	+1 (617) 555-1212	work phone
email address	an electronic mail address	MeMyselfAndI@gmail.com	personal email address
internet address	a web address, Uniform Resource Locator	www.DaoustAssociates.com	web address
name-person	a name of a human being	Norman Joseph Daoust	spouse name
name-organization	a name of an organization	IIBA Greater Boston Chapter	chapter name
name-other	a name of something other than a person or organization	Monthly Chapter Meeting	meeting name
identifier	a value that uniquely distinguishes one occurrence from all others	1321 5823 0583 2910	credit card number
specification text	a textual specification that can be parsed into distinct data elements	2nd Thursday of the month	meeting schedule
text	human readable text, frequently a description of something	When, in the course of human events, it becomes necessary …	product description
encoded data	a file, document, etc. that typically requires special software to unencode or view it	ND-picture.jpg	picture

Some of these datatypes (e.g. person name, address) are compound: they can be broken down further. We've found that utilizing these compound datatypes keeps our analysis models smaller, simpler, and easier to understand for subject matter experts. However, don't neglect to document the structure of these compound datatypes in detail as part of your model!

Implementation Considerations for Analysis Datatypes

If you're concerned about design and implementation, you'll want to read our Implementation Considerations for Analysis Datatypes.

Mappings to Database Managment System (DBMS) Data Types

In case your software developers or database administrators are wondering how to translate these analysis datatypes into the datatypes of your specific database management system, several experts have generously provided that information for you:

• SQL Server
• Access

If we don't list a mapping to your favorite database management system (DBMS) or programming language, please contact us.

For those that don't understand the need for analysis datatypes in analysis data models and UML class models and only think in terms of a particular database management system's datatypes, I suggest referring to The Zachman Framework, in particular, the What column of the Business row (sometimes known as row 2, column 2). That's the type of model for which these analysis datatypes is applicable.

Acknowledgements

Thanks to the following people for their contributions:

• Andrew Novick of Novick Software, Inc.
• Ed Landale