What is Business Intelligence? A broad category of applications, technologies, and processes for gathering, storing, accessing, and analyzing data to help business users make better decisions.
History of BI: Past, Present, & Future “Knowing the past is important because it provides context and helps in understanding many of today’s practices.” Hugh Watson, 2009
BI definition, frameworks, and concepts History of BI BI origins BI definition, frameworks, and concepts BI trends & resources
BI Origins Much of the early work was done at MIT & Harvard In his dissertation work (1967), Morton built, implemented, and tested a “management decision system” to support the coordination of production planning for laundry equipment Michael Scott Morton
BI Origins Many “decision support systems” were developed from late 1960s to mid 1970s Sprague’s framework Alter’s typology Power’s taxonomy Ralph Sprague, Jr. Steve Alter Daniel J. Power
Sprague’s Framework Sprague, R. & Watson, H. 1979. “Bit by Bit: Toward Decision Support Systems,” California Management Review, p. 64.
Alter’s Typology Alter, S. 1975. “A study of computer aided decision making in organizations,” Doctoral Dissertation: MIT, p. 5.
Power’s Taxonomy Communications Data Document Knowledge Model Source: http://dssresources.com
Power’s Taxonomy Communications Data Support collaboration and communication Audio conferencing Bulleting boards Data Emphasize access to and manipulation of a time-series data stored in a data warehouse Dashboards/scorecards Business reports Source: http://dssresources.com
What is a data warehouse? A database designed to support decision making in organizations. It is batch updated and structured for rapid online queries and managerial summaries. Data warehouses contain large amounts of data. A data warehouse is a subject-oriented, integrated, time-variant collection of data in support of management's decision making process
Power’s Taxonomy Document Knowledge Support the search and retrieval of unstructured documents Transcribed conversations Memos Knowledge Incorporate knowledge, experience, and judgment of experts in a particular domain and recommend courses of action Expert system Source: http://dssresources.com
Power’s Taxonomy Model Emphasize access to and manipulation of a model, for example, statistical, financial, optimization and/or simulation Revenue management Production planning Source: http://dssresources.com
BI Origins— The more things change… Internet and web-based applications changed how people search for information Encyclopedia in physical library vs. Google search Yet, much remains the same Critical success factors (1979) vs. Key performance indicators (1997) vs. Dashboards/Scorecards (now) Iterative, evolutionary design (1980) vs. Agile design (now) Names and some nuances may change, but many of the fundamental concepts are the same
What is Business Intelligence? A broad category of applications, technologies, and processes for gathering, storing, accessing, and analyzing data to help business users make better decisions. The term was coined in 1989 by Howard Dresner (Gatner analyst)
Three BI targets A single BI application Enterprise-wide BI A point solution for a departmental need A data mart is created to provide data Enterprise-wide BI The creation of infrastructure supporting current and future BI needs A data warehouse is created Organizational BI transformation BI is used to fundamentally change how a company competes in the marketplace Supports a new business model and enables the business strategy 3AM = Its divisions operated independently and had separate decision support platforms Harrah’s = created a customer loyalty program encouraging cross-casino play. It needed a BI infrastructure collecting data across various customer touchpoints (slot machines). Using this data, Harrah’s became an industry leader because of its ability to reward loyal customers and reach out to them in a personal way
How is a data mart different than a data warehouse? A data mart is a simple form of a data warehouse that is focused on a single subject (or functional area), such as Sales or Finance or Marketing
A generic BI environment— “Getting the data in” and “Accessing the data”
Source systems “Getting the data in” All systems that provide BI with data Single-source vs. enterprise-wide Type of source system is chosen based on the business need motivating BI. Yet, a variety of systems can be used ERP Operational systems Customer demographic data
Source systems “Getting the data in” Typically use different platforms (IBM & Oracle) and store data in different formats (relational vs. hierarchical) Profile data to verify documentation quality Additional issues What is the best source system? How granular is the data?
Data integration “Getting the data in” Extract, transform, and load (ETL) Extraction can be performed by hand-written code (SQL), yet it is usually performed by commercial software Transformation based on business rules Use of Male and Female to denote gender instead of 0 and 1 or M and F Loading occurs during a “window” The window is narrower (real-time)
Storing data—Architecture “Getting the data in” Independent data marts Do not provide a “single version of truth” Enterprise data warehouse Dependent marts are created in iterative manner Provide a “single version of truth” Centralized Data mart bus (Kimball Bus Architecture) Dimensional data model Federated Data integration through shared keys
Storing data—Architecture “Getting the data in” Enterprise data warehouse = Data mart bus > Federated > Independent data marts (Ariyachandra & Watson 2006)
Who are the users? “Accessing the data” Information producers IT developers Analysts Information consumers Managers and executives Front line workers, suppliers, customers, regulators
What are the applications? “Accessing the data” SQL IT developers & analysts Business reports Front line workers Excel (most popular) Dashboards/Scorecards Data visualization Predictive analytics
MicroStrategy Dashboard
Metadata “Accessing the data” Data about the data (important!) IT personnel who “get data in” need to know What data is stored where Table and attribute names Users need metadata to support “getting data out,” including Timeliness and data quality Who has access to specific data and reports Do not use Excel!
Data quality “Accessing the data” Major problem in most companies Data profiling is a start but correcting at the source is a desired solution “There isn't a “quick fix” for data quality problems” – Think of Nestle! Data must be accurate enough Consistently well defined and used Necessary data must be available and timely Easily accessible, understandable, and usable
Governance— Multi-level & cross functional Strategic level VPs often provide strategic direction that is aligned with company objectives Tactical level Directors ensure projects are on time and schedule Operational level Users address metadata and quality issues
BI benefits
Note to Carolina: Stop here! Finish material on Wednesday!