Topic 1.1 - Data Analysis & Decision Making¶

Unit 1 - Data Analysis for Decision Making¶

1. Data Analysis: Data to Information¶

Data:

Unorganized raw facts
Any sequence of one or more symbols without context or meaning

Information:

Processed and organized data
Presented within a given context that is useful to humans
Given meaning by specific act of interpretation involving human judgement

2. Data Analysis & Decision Making¶

How data support decision making:

Demonstration: Data Supporting Decision Making

Citation: Jean Francois Puget:

The purpose of an analytics solution is to enable decisions based on data.

It does not necessarily mean that such solutions must compute decisions.

It means that whatever their output, they must support decision making.

In short, analytics is a journey from data to decisions and actions, via the use of many different techniques and tools.

Unit 2 - Data Analytical Process: Data Wrangling & Analyzing & Reporting¶

Data Analytical Process Overview: An Iterative & Repetitive Process:

1. Data Wrangling¶

Data wrangling is everything you do except analyse the data, so that you can analyse the data.

Demonstration: Time & Efforts in Data Analysis Stages

Citation: Gil Press:

Data scientists spend 60% of their time on cleaning and organizing data

Collecting data sets comes second at 19% of their time

80% of their time on preparing and managing data for analysis

Demonstration: Data Analytical Process

Demonstration: Data Visualization as an Iterative Process

Demonstration: Enterprise Data Pipeline

2. Data Analyzing & Reporting¶

(1) Data Analysis Trifecta: Data + Intent + Tools & Technologies¶

(a) Data

Data analysis project involve sourcing, exploring, and evaluating data:

Data analytics project may not begin with a pre-supplied dataset
Efforts must be made in exercising judgement over:
Which data is reliable & relevant
What weighting it should be given when making decision

(b) Intent

Data Analytical Intent:

Gain: Gain insight about what is happening in the world around us
Understand: Understand how we can influence what happens
See: See trends or patterns in activity and behavior
Predict: Predict events and outcomes
Set: Set goals and measure events and outcomes against those goals
Take: Take actions that lead to desired outcomes and achievement of goals

Two important feature of data analytical process:

Data analysis requires being familiar with the specific context of your data analysis project:
- Start a new analytics project:
  - Understand the desired business outcome and how it will be measured
  - Understand the decisions in scope that influence this outcome
  - Access the sponsorship who can assist you with this understanding
- Manage sustained project engagement: prototyping, iterating, and getting feedback from users
  - Help develop a useful, relevant and more likely to be adopted solution
  - Ensures that the solution aligns with real user needs
- Use specific examples
  - Illustrate the decision and context within which it is to be made
  - Reveal additional factors, goals, and strategies that also influence the decision
  - Be wary of over-solving the example instead of developing generalized approach
Data analysis is a time-consuming and costly process
- Producing a fully automated analytical solution is costly
  - Require deep understanding of the decision making context
  - Require consensus among all the incumbent decision makers
- The analytical output is expected to be recommendations not facts or binding mandate
- Not as simple or effortless as the buzzwords and marketing in data analysis, for example:
  - Rules-based engine,
  - Real-time factors and triggers,
  - Just give us access to your information...

(c) Tools & Technologies

Tools & technologies need to be guided by specific purpose, to produce something useful.

Demonstration: Computational Thinking

Citation: Denning, Peter J & Tedre, Matti (2021):

Computational Thinking: A Disciplinary Perspective, Informatics in Education:

Computational thinking is the mental skills and practices for designing computations that get computers to do jobs for us, and for explaining and interpreting the world in terms of information processes.

Citation: www.freecodecamp.org:

Generalized Problem-solving skills are almost unanimously the most important qualification that employers look for.

Understand -> Plan -> Divide & Combine

Relevant to other activities, not just computer programming, e.g., research projects, other disciplines.

(2) Analytical Levels¶

Analytical Levels Concepts:

Analysts build solutions to aids decision maker from data to insight, to decision, and to action
Analytical levels describes:
How far the analyst pursues this journey
How many works left to decision makes to complete this journey

Analytical Levels Definition:

(1) Descriptive:
- What happened or is happening?
- Perspective on the data as is
(2) Diagnostic:
- Why did it happen?
- Insight about patterns or relationships
(3) Predictive:
- What will happen?
- Predict future events based on patterns or relationships
(4) Prescriptive:
- What should we do?
- Recommendations or actions that influence the future

Analytical Levels Comments:

Four analytical levels are overlapping rather than separate
The different levels depend on:
- What is automated analytics work (automation is imitating human judgement)
- What is left to additional human effort and judgement
Data analysis is the ongoing process of turning data into decisions (pushing arrow right)
- Great deal of human effort and judgement is required
- Either by an analyst, or collective wisdom captured into rules and calculations

Demonstration: Computational Thinking

In the context of analyzing sales:

(1) Descriptive - What happened?

Summarizes past sales, e.g., revenue, units sold
Breaks down by time, product, region, customer
Shows key KPIs and trends with charts

(2) Diagnostic - Why did it happen?

Analyzes sales by channel or location
Identifies causes of growth or decline
Explores customer behavior, returns, or inventory issues

(3) Predictive - What will happen?

Forecasts future sales using historical data
Uses models, e.g., time series, regression
Helps plan for demand, promotions, and budgets

(4) Prescriptive - What should we do?

Recommends actions based on predictions
Plans inventory, production, and staffing
Aims to reduce costs and optimize operations