School of natural science

Department of mathematics and computer science

Name: Tendai ashineki shonhiwa

Reg number: m203212

Program: BSc hon in computer science

Course title and code: business intelligence (hcs4113)

 1. Explain in detail the reasons for utilising models in decision making as an important
aspect in Business Intelligence. [20]

In the context of Business Intelligence (BI), an utilising model can refer to the approach that an
organization takes to adopt, utilize, and gain success from its BI systems. This model is often driven by
the specific needs of the organization and can vary widely based on factors such as the organization’s
size, industry, and strategic goals. Utilizing models in decision making is a crucial aspect of Business
Intelligence (BI) for several reasons:

Data-Driven Decisions:
The ultimate goal of BI is to help organizations make more data-driven decisions. By providing a
comprehensive view of an organization’s data, BI tools can help to drive change, eliminate inefficiencies,
and quickly adapt to market or supply changes. Models help organizations make decisions based on solid
data rather than intuition or guesswork. This leads to more accurate and reliable outcomes. For example, a
retail company might use BI models to analyze sales data and identify trends, which can then inform
decisions about inventory management or marketing strategies.

Real-Time Insights:
Models provide real-time insights, allowing companies to respond swiftly to market changes and internal
performance metrics. For instance, a manufacturing company might use BI models to monitor production
data in real-time, enabling it to quickly identify and address any issues that arise.

Predictive Analysis:
Models can also be used to predict future trends based on historical data. This can help organizations
anticipate future challenges and opportunities, and make proactive decisions accordingly. For example, a
financial services company might use BI models to predict future market trends, which can then inform
its investment strategies.

Improved Collaboration and Communication:

By providing a centralized source of truth, models can improve collaboration and communication within
an organization1. Teams can work together more effectively when they have access to the same data and
insights.
Efficiency and Productivity:
By automating data collection and analysis, models can significantly increase efficiency and
productivity. Employees can spend less time on manual data tasks and more time on strategic decision-
making.

Understanding Business Requirements:

Understanding business requirements is a critical step in data modelling for business intelligence. It
involves identifying the key goals, objectives, and challenges of the business to determine what data is
needed and how it should be modelled. This process requires collaboration between business stakeholders
and data analysts to ensure that the resulting data model aligns with the needs of the organization.

In conclusion, utilizing models in decision-making is an important aspect of Business Intelligence. It’s

about leveraging data to make informed, strategic decisions. It’s about transforming raw data into
actionable insights that can drive business success.

2 a) with the aid of clearly labelled diagrams describe the following data warehouse
architectures:

i. Basic data warehouse architecture [4]

A basic data warehouse architecture consists of the following components:

Operational System:
This is a system used to process the day-to-day transactions of an organization.
Flat Files: A system of files in which transactional data is stored. Every file in the system must have a
different name.

Meta Data:
A set of data that defines and gives information about other data. It summarizes necessary information
about data, which can make finding and working with particular instances of data easier.

Lightly and Highly Summarized Data:

The area of the data warehouse saves all the predefined lightly and highly summarized (aggregated) data
generated by the warehouse manager. The goals of the summarized information are to speed up query
performance.

End-User Access Tools:

ii. Data warehouse with staging area and data marts [4]

A data warehouse with a staging area and data marts involves a more complex architecture. It consists of
the following

Staging Area:
The staging area is a temporary space where data from multiple sources is gathered before it undergoes
the transformation process. The data here is raw, unprocessed, and often in different formats, representing
a myriad of sources from where it originates. The primary purpose of a staging area is to ensure a smooth
and efficient ETL process. It does this by cleaning and organizing the data, removing redundancies, and
ensuring data consistency.

ETL Process:
ETL stands for Extract, Transform, Load. This process moves data from its raw form, reshapes it, and
finally loads it into the data warehouse.

Data Marts:
Data marts are specialized sections of the warehouse where data is stored in a way that’s specific to the
needs of different teams or departments in a business. They are smaller than the main data warehouse and
can be used to provide data to specific business lines or teams within the organization.

Data Warehouse:
The data warehouse is the central repository of data, which has been cleaned, transformed, and cataloged
so that it can be used for business analysis and reporting.

This architecture allows for efficient data management and retrieval. The staging area ensures that the
data is clean and consistent before it enters the data warehouse, and the data marts allow for efficient
access to data for specific business needs. The figure below illustrates an example where purchasing,
sales, and stocks are separated. In this example, a financial analyst wants to analyze historical data for
purchases and sales or mine historical information to make predictions about customer behavior.

b) “Dependant data marts can be used to bring sanity to an organisation’s data warehouse
user community”. With reference to an organisation of your choice, discuss the reasons
why an organisation may be driven to create dependent data marts, thereby justifying the
above assertion. [6]

Dependent data marts are subsets of a data warehouse that are developed to serve specific reporting and
analytical needs. They are often updated from the data warehouse, benefiting from its data integration,
data quality, and consistency. This allows for the centralization and preservation of all data in a single
source of truth. A hypothetical organization, “Muchiso Inc ”, that operates in the technology sector.
Muchiso Inc may be driven to create dependent data marts for several reasons:

Performance:
If the performance of Muchiso Inc data warehouse becomes an issue, building one or two
dependent data marts can solve the problem. This is because data marts contain a smaller, more
focused subset of data, which can be queried and analyzed more quickly than the entire data
warehouse.

Security:
By putting data outside the data warehouse into independent data marts, each department at
Muchiso Inc owns their data and has complete control over it. This can help to protect sensitive
information and ensure that only authorized personnel have access to specific data.

Relevance to Use Cases:

Data marts can be tailored to the specific needs of different teams within Muchiso Inc. For
example, the sales team might have a data mart that contains detailed sales data, while the
marketing team might have a data mart that focuses on customer demographics and campaign
performance.

Accessibility:
Data marts allow Muchiso Inc to expose more people to data without overwhelming them. They
provide a smaller schema with only the relevant tables for the group, making it easier for
individuals and teams to navigate and understand the data.

Customized Architecture for Different Use Cases:

 Aggregations, metric calculations, and personally identifiable information (PII) can all be handled
individually for teams. This allows each team at Muchiso Inc to work with data in a way that best
suits their specific needs.

Maintainability:
Having the data monitored by team leads makes it easier to identify data issues3. This can help to
ensure the accuracy and reliability of the data in Muchiso Inc data marts.

KPI Tracking:
Dependent data marts are ideal places for building and tracking Key Performance Indicators
(KPIs) over a long period of time. This can provide valuable insights into Muchiso Inc
performance and help to guide strategic decision-making.

In conclusion, dependent data marts can indeed bring sanity to an organization’s data warehouse
user community. They provide a more manageable, secure, and user-friendly way to access and
analyze data, thereby enhancing the value that Muchiso Inc can derive from its data warehouse.

c) Data mining can be used to discover patterns and relationships in information from data
warehouses and data marts. Discuss using examples three different types of results that can
be obtained from data mining operations. [6]

Data mining is a powerful tool that can extract valuable insights from large datasets. Here are three types
of results that can be obtained from data mining operations:

Pattern Discovery:

Data mining can identify recurring patterns within a dataset. For example, a retail company might use
data mining to find patterns in sales data that indicate which products are frequently bought together. This
information could then be used to develop effective marketing strategies, such as product bundling or
targeted advertising.

Anomaly Detection:

Data mining can also be used to detect anomalies, or unusual data points, within a dataset. For instance, a
credit card company might use data mining to identify unusual spending patterns that could indicate
fraudulent activity. By detecting these anomalies quickly, the company can take immediate action to
prevent further fraudulent transactions.

Dependency Modeling:
 Data mining can discover relationships between different variables within a dataset. This often involves
regression analysis. For example, an automobile manufacturer might use data mining to understand the
relationship between different vehicle features (like engine size, weight, and fuel efficiency) and the
vehicle's price. This information could then be used to guide the design and pricing of new vehicle
models.