Transcript Introduction to Business Intelligence

Decision Support for
Management
Data+ Models+Intuition
• Decision support systems –use data and
models to support management decision
making in different ways.
• Nowadays, what used to be called decision
support systems often comes under the
umbrella of business intelligence.
• Analytics describes the application of
mathematical techniques to organisational
operations.
Collecting Data
• Organisations need to collect data in order
to be able to understand and improve their
business.
• They need models (from analytics) to
interpret this data.
April 8, 2016
3
Example:Looking at data
Harmful Algal Blooms
• www.marine.ie
• Click on the data tab
• Click on HABS search
• And tell me whether Donegal bay and Sligo
are allowed to sell fish/shellfish or not.
• [in addition to automatic notification by email, fax and SMS]
Modeling and Data
• A model is a selective abstraction of reality.
– Selective
• We choose which bits to put in the model.
– Abstraction
• The model is not reality it is a simplification.
5
Modelling
• A model is way of representing a part of
the environment.
• Trade-off between the simplification and
the representation of reality.
– Advantages of simple models.
– Disadvantages of simple models.
6
Types of Models
• Mental Models
• Visual
– Also called analogue.
• Physical/Scale
– Also called iconic.
• Mathematical
– Also called quantitative.
7
Benefits of a Model
•
•
•
•
Can compress time.
Can manipulate easily.
Can do trial and error calculations.
Can model risk and uncertainty.
8
Benefits of Producing Models
• Having a model to use is beneficial but the
process of producing a model is equally if not
more beneficial.
• Helps you to understand the problem.
– Need to be explicit about your goals.
– Need to quantify the variables which affect the goals.
– Need to identify constraints and relationships between
variables.
– Facilitates communication and understanding.
9
Example: Simulation Models
• http://www.bized.co.uk/virtual/economy/
model/
• http://www.xjtek.com/anylogic/demo_mo
dels/?simulation_method=System++Dynam
ics
Business Intelligence
Overview
BI involves acquiring data and
information from a wide variety of
sources and utilising them in
decision-making.
• in-depth analysis of company data for
better decision-making.
• Models are used to analyse data.
Business Intelligence
• Business intelligence (BI) simplifies
information discovery and analysis, making
it possible for decision-makers at all levels
of an organization to more easily access,
understand, analyze, collaborate, and act
on information, anytime and anywhere.
(Microsoft)
• The technology and processes that make
this analysis possible take unwieldy
collections of information and translate
them into organized, readily-accessible,
human-readable compilations of data.
What can companies do with BI?
Track
• their own operations
• customers’ activity patterns
• industry trends.
• fact-based assessments help companies
work toward specific goals with confidence.
Data is
1. Gathered from relevant sources
2. Filtered, and stored
3. Analysed and arranged into meaningful
patterns using different tools .
4. Business intelligence is the knowledge
gained from that data analysis.
Overview of Business Intelligence
Analytical
tools
Data
Sources
Data
Warehouse
Data
visualisation
OLAP
Data
Mining
Analysis
Results
Data
visualisation
....
From Turban, Aronson and Liang
Some Questions
• Where does the data come from?
• How can we decide what data is important?
• How can data from different sources be
joined together (consolidated and
integrated) securely?
• How can data be analysed?
• How can these analyses be viewed?
Where does the data come
from?
• Data can be collected manually or
automatically.
– Transaction data e.g. supermarket checkout,
bank withdrawal
– Time studies, questionnaire, observation notes
– Physical sensors e.g. temperature of a rooms in
a house
– Sensors, scanners, bar codes
How can we decide what data is
important?
• Depends what our goals are, the functional
area(e.g. Sales, HR, marketing..) and what
processes we are looking at..
Balanced scorecard
Critical success factors
Key performance indicators
Human
resources
•employee
•organizational
• departmental
measures
Sales and marketing
• products
Functional
• customers
• demographics
Areas
• promotions
• sales force
Finance
• order type
• currency standards
• account information
• industry trends
Operations
management
•assembly speed
•warehouse stock
•manufacturer and
supplier cost
•shift productivity
Data Quality is also important
Garbage in..... Garbage out
• Contextual – relevance, value, timeliness
completeness, amount
• Intrinsic – accuracy, objectivity, believability,
reputation
• Accessibility DQ – ease of access,security
• Representation DQ – interpretability, ease of
understanding, concise, consistent
representation.
Example : ecological intelligence?
What information can we access on
www.goodguide.com
What do you think the goals of someone using
this website might be?
What type of data do you think has been
analysed to give this information?
Where did the data come from?
How reliable is the data?
Can you find out how it was analysed?
Example: What is ecological
intelligence?
• a vast, shared network of detailed
information regarding the full social and
ecological impact of products. Consumers
will be able to use an array of new wireless
and web-based technologies to instantly
tap into this network to find product
information, even at the point of purchase.
Example: Ecological Intelligence
How is the data analysed?
• Industrial ecologists and engineers
deconstruct the ingredients and processes
that go into any manufactured object and
do a Life Cycle Assessment, or LCA. This
allows them to track a product’s precise
social, health and ecological effects from
production to final disposal.
Example Data Warehouse Technology : Microsoft
How can data from different sources be joined
together (consolidated and integrated) securely?
• SQL Server provides a comprehensive and
scalable data warehouse platform
• organizations build large-scale enterprise
data warehouses that can consolidate data
from multiple disparate systems into a
single, secure, manageable solution.
What is a Data Warehouse?
A data repository that makes operational and
other data accessible in a form that is
readily acceptable for decision support and
other user applications.
Note: A data warehouse is not another word
for a database. The specific purpose of a
data warehouse is to support decisions not
operations.
28
Data 3
Data warehouses vs operational
databases
•
•
•
•
•
•
•
an operational database is normalised. Each data item is only held
once.
databases have very fast insert/update performance because only a
small amount of data in those tables is affected each time a
transaction is processed.
Older data may be periodically purged from operational systems to
improve performance.
Data warehouses are optimized for speed of data retrieval.
data in data warehouses may be stored using a dimension-based
model.
To speed data retrieval, data warehouse data are often stored
multiple times.
Data may be held in the data warehouse even after the data has
been removed from the operational systems.
How is the data analysed?
Analytics techniques – types of model
–
–
–
–
–
–
Simulation
Decision analysis
Statistics : averages, correlations,
Linear programming: optimisation
Queuing theory: “waiting line”analysis
Network analysis: Maximise flow through a
network e.g. A supply chain
– Multi-criteria decision making: scoring models
Example – Microsoft OLAP
How can data be analysed?
• Microsoft Online Analytical Processing
(OLAP) makes it quick and easy to perform
ad-hoc queries and analysis of large
amounts of complex data across all aspects
of your business.
Example – Microsoft OLAP
Microsoft OLAP is used to report
on...
•
•
•
•
•
•
sales
marketing
management issues
business process management
budgeting and forecasting,
financial issues etc..
What is OLAP?
• OLAP enables you to look at and access
data in different ways (3-d data cubes) ,
drill down, view summarised data, make
calculations on the fly etc.
• http://spatialolap.scg.ulaval.ca/Examples.a
sp
• http://www.census.gov
What is Data Mining?
• Data mining is a capability to support the
recognition of previously unknown but
potentially useful relationships within large
databases/ data warehouses.
• Basically software to analyse data and spot
patterns.
Visualising Data
•
•
•
•
•
•
Digital images- These can be still or animated.
Maps e.g. Geographic Information Systems
Multidimensions - (OLAP)
Tables and graphs
Virtual reality
Dashboards
A Table
A Chart
€8 0 0 . 0 0
€7 0 0 . 0 0
€6 0 0 . 0 0
€5 0 0 . 0 0
Qt r 4
€4 0 0 . 0 0
Qt r 3
Qt r 2
€3 0 0 . 0 0
€2 0 0 . 0 0
€10 0 . 0 0
€0 . 0 0
A Behan
B Collin s
C Daly
G En n is
A Nolan
A Sheils
T T immon s
J Wyn n e
Dashboards
Taken from http://gbr.pepperdine.edu/034/bis.html
multiple, synchronized chart types
A visualization with multiple displays showing a Supplier scorecard in conjunction with a
geographical display.
Return to Document
Summary
• Decision support involves data and models
• BI involves acquiring data and information
from a wide variety of sources and utilising
them in decision-making. Data is
–
–
–
–
Gathered, selected
Consolidated and integrated -> data warehouse
Analysed in different ways (analytic techniques)
Results are Visualised
We need to Understand
•
•
•
•
•
•
•
Data issues – data quality
Where data comes from
How data is stored: data warehouses
How data is analysed
Tools to do this.
Limitations of the computer
Our own blind spots (if this is possible)!
References
Advanced Analytics- Information Week 2010
(analytics.informationweek.com)
Competing on Analytics - Thomas Davenport
Harvard Business Review Jan 2006
In search of Clarity - Economist intelligence
unit 2007 (available from sap)
What is Business Intelligence (sap)