Decision-Making and DIKW Model - Week 12

Week 12

• The DIKW model shows how data progresses through knowledge.
• Analytics leads to decisions, which are based on wisdom.
• Decisions involve considering consequences and stakeholders' responses.
• Traditional decision-making relies on human judgment, experience, and expertise.
• Decision types span managerial, tactical, and strategic levels.

Decision Types

• Decision-making complexity increases with management level.
• Operational decisions focus on daily operations, involving data input from transaction processing systems.
• Managerial/tactical decisions concern resource allocation and short-term strategies.
• Strategic decisions address long-term organizational issues, needing high-level knowledge.
• Strategic decisions are usually made in groups.

Data-driven vs Judgment-driven

• Manager types demonstrate varying use of analytics and judgment.
• High data-driven types (Type A) heavily use analytics.
• High judgment-driven types (Type D) rely heavily on judgment.
• Intermediate types (Type B and C) use a balanced approach.

Data Ethics (PAPA Framework)

• Privacy, accessibility, accuracy, and property are ethical considerations.
• Privacy concerns personal information, what people can reveal.
• Accessibility questions who has access and under what conditions?
• Accuracy deals with data reliability and accountability.

Extended PAPA Frameworks

• Newer ethical challenges created by data types (big data, etc), different types of analysis, and social media.
• These require more considerations in Extended PAPA Frameworks.

Skills for Business Analysts

• Soft skills include communication, critical thinking, problem-solving, collaboration, adaptability, negotiation, and time management.
• Hard skills include data analysis (Excel, SQL, statistical software), data visualizations (Tableau, Power BI), programming (Python or R), technical writing, and database management.

Week 11 - Clustering

• Clustering groups data based on similarities.
• It lacks a target variable.
• Examples include grouping customers based on demographics and buying behavior.
• Clusters are often difficult to identify and label.
• K-means is a referenced clustering algorithm.
• k values are determined by the analyst.
• The algorithm assigns data points to the nearest cluster center.
• It re-computes centers until the cluster assignment stabilizes.

Association Rule Mining

• Finds relationships (co-occurrences) amongst data variables.
• Often used to analyze market basket data.
• Apriori algorithm frequently identifies subsets that occur together.
• This is crucial for promoting related products.

Data Mining vs Statistics

• Data mining uses a wider range of data than statistics.
• Statistical analysis typically involves testing hypotheses from smaller samples.
• Data mining frequently involves analysis of large amounts of data.

Week 9 - Data Mining

• Data mining is the process of extracting useful information from large datasets.
• It uses statistical and algorithmic methods to discover patterns.
• CRM management, financial analysis, and medical applications are examples.
• It can provide insights to improve operations.

Main Types of Data Mining Methods

• Prediction involves predicting future occurrences.
• Classification involves sorting data into categories.
• Clustering involves grouping similar data points.
• Associations identify co-occurring variables.
• Time-series analysis examines trends over time.

Classification vs Regression

• Classification predicts categorical outputs (e.g., "Yes" or "No").
• Regression predicts numerical outputs (e.g., temperature).

Week 6 - Big Data

• Big data refers to large datasets that traditional methods cannot handle.
• It has four characteristics: Volume, Variety, Velocity, and Veracity.
• Big Data can be used in diverse ways, e.g., advertising, healthcare, and retail.
• Sources of big data include many kinds of digital archives, documents, and media.

Data Warehouses and Data Marts

• Data warehouses store an organisation's entire enterprise data.
• Data marts focus on specific departments or smaller business units.
• Data warehouses include processed or prepared data.
• Data marts utilize data from the warehouse but may not process data as deeply.
• Data lakes are locations that house raw and unprocessed data.

Week 4 & 5 - Databases

• Databases store and organize linked records.
• Tables hold the data.
• Relationships link different tables based on associated data values.
• Tables have primary key fields that identify each record uniquely.
• Foreign key fields in another table link records in different tables.
• Metadata describes the data.

Week 4 & 5 - Data Modelling

• Data modeling involves visually representing database structure.
• Relational databases organize data in tables.
• Entity-relationship diagrams (ERDs) show how entities (types of data or things to be represented) relate to one another.

Week 4 & 5 - ERD Notations

• Rectangular boxes represent entities (e.g., Customers, Employees, or Sales).
• Ellipses represent attributes (e.g., customer name or product quantity or price).
• Diamonds represent relationships.
• Cardinality represents the relationship’s strength (One to Many, etc).

Week 3 - Business Process Management (BPM)

• BPM is a methodical approach to improving or defining business processes.
• It includes four key stages: Modeling the current process; creating system components to support the updated process; implementing the updated business process; and monitoring its effectiveness (using policies and ongoing assessments)
• BPM helps streamline workflows, improve efficiency, reduce costs, and ensure improved customer value.

Week 3 - Varied types of BPM

• Functional processes may involve a single department or function.
• Inter-functional processes involve more than one department in a company.
• Inter-organizational processes require collaboration beyond organizational boundaries.

BPM Notation

• BPM notation tools capture processes.
• Using symbols to indicate different elements of a process helps convey the steps and logic.

Week 2 - Business Processes

• Processes are structured networks of activities, encompassing resources, facilities, and data.
• Information systems are part of business processes.
• Transforming raw information into a useful output is fundamental to any business process.
• Increased use of technology has a direct impact on how business processes operate (e.g., increased automation).

Week 1 - Business Analytics

• Business Analytics (BA) is a set of procedures and methods for solving business problems through data analysis.
• Data analysis processes include planning, data pre-processing, modeling, and follow-up.
• Statistical model results can help provide meaningful, quantitative insights into a business issue.
• BA uses tools like SQL, Python, R Studio, and Tableau.

Types of Decision Modeling

• Deterministic models utilize known data.
• Probabilistic models incorporate uncertainty using probability distributions.

Explore the intricacies of decision-making processes and the DIKW model in this quiz. Understand how data evolves into knowledge and leads to informed decisions at various managerial levels. Delve into the contrasts between data-driven and judgment-driven decision-making styles.