Decoding Your Data

 Unraveling the World of Data and Its Insights

Learning Outcome

Learners should know :

These operations rely on data structure and classification principles explored today

Imagine a company records every transaction.

Millions of Entries

Time

Amount

Location

Date

But until someone investigates patterns,
groups results,
compares timelines

The number exist

The Answer exist

The insight was hidden in plain sight.

Data holds answers, but only structured thinking reveals them.

Before advanced analytics or predictive systems, we must establish:

• What exactly qualifies as data?

• How does meaning emerge from raw records?

• How does classification affect analysis?

• What structured process ensures reliable conclusions?

We begin by examining the foundational distinction between data and information.

 Data

Data consists of raw observations, measurements, or recordings collected from events, systems, or environments.

This could represent units sold, revenue, distance, or temperature.

Without context, interpretation is impossible.

Technical Perspective

Data represents variables — attributes that describe entities.

Information

Information is the result of processing, organizing, structuring, or analyzing data to extract meaning.

Characteristics

Imagine Sales increased by 8% after launching a promotional campaign

Information answers a specific question.

Types of Data

Understanding type determines:

Qualitative Data

Descriptive data representing characteristics or categories.

Example

 Quantitative Data

Numeric data representing measurable quantities.

Key Characteristics:

Supports arithmetic operations

Enables statistical modeling

Can be continuous or discrete

Example

These is subtypes of qualitative data

Country Names

Product types

Example:

Equality
comparision

Frequency
calculation

Allowed operation

You should not treat ordinal data like numerical data

These is subtypes of quantitative data

Mathematically represented as distinct points.

Ratio Data

Equal spacing

• True zero exists

Supports full mathematical operations

Addition

Subtraction

Multipication

Division

• Proportional statements

• Meaningful comparisons

Ratio data allows:

Example:

• Revenue

• Distance

• Weight

Interval Data

• Equal spacing between values

• Zero is not Absolute
• Zero ≠ Nothing

• Differences are meaningful

• Ratios are not meaningful

Example:

• Temperature in Celsius

Zero does not represent absence of temperature.

These is subtypes of quantitative data

Example:

Weight

Temperature

Time duration

Structured Data

Data organized into predefined schema, typically rows and columns.

Characteristics:

• Fixed format

• Clearly defined data types

• Easily searchable and queryable

• Efficient for relational storage

Advantages:

• Fast querying

• Clear relationships

• High data integrity

Semi-Structured Data

Characteristics:

• Flexible schema

• Uses tags or key-value pairs

• Hierarchical structure possible

Definition:
Data that does not conform to strict tabular structure but contains organizational markers.

JSON

Log Files

XML

Uses

Unstructured Data

Characteristics:

• Large volume

• Difficult to analyze using traditional techniques

• Often requires advanced processing (text mining, image processing)

Definition:
Data without predefined organization or schema.

Images

Audio

VIdeos

Free-text documents

Uses:

Data Analysis: Conceptual Framework

Data analysis is a structured methodology for converting observations into insight.

Analysis reduces uncertainty.

 Steps of Data Analysis

Step 1 – Data Collection

• Identify relevant sources

• Ensure reliability

• Avoid biased sampling

Poor collection leads to flawed conclusions.

 Steps of Data Analysis

Step 3 – Analysis

Techniques include:

• Descriptive statistics

• Correlation analysis

• Comparative analysis

• Trend analysis

• Predictive modeling

Choice of method depends on data type
and objective.

 Steps of Data Analysis

Step 4 – Interpretation

Involves:

• Evaluating statistical significance

• Assessing practical impact

• Communicating implications

• Recommending actions

Insight must align with business
or research goals.and objective.

 Qualitative Analysis

• Thematic coding

• Sentiment evaluation

• Content analysis

Exploring

behaviors

Investigating open-ended responses

Understanding
perceptions

Focuses on understanding meaning.

Method Include

Used When

 Quantitative Analysis

Forecasting

Regression analysis

Measuring performance

Validating assumptions

Predicting outcomes

Focuses on measurable relationships.

Method Include

Used When

Hypothesis testing

Statistical testing

 Mixed Methods

Combines qualitative depth with quantitative validation.

Advantages:

• Broader understanding
  combines the strenght of both methods

• Balanced insight
  Provides a fuller picture of the research topic

• Reduced bias
  offsets the limitation of each individual approach

The Data Science Lifecycle

Stage 1 – Problem Definition

A poorly defined problem leads to irrelevant analysis.

Clearly define:

Objective

Constraints

Success metrics

Stage 3 – Analysis & Modeling

Apply statistical or machine learning techniques

Train and validate models

Evaluate performance metrics

Iterative Nature

New findings may:

• Refine the original problem

• Require new data

• Adjust modeling approach

The lifecycle is not linear.

Iteration improves precision and reliability.

Summary

Which classification determines whether mathematical ratios are meaningful?

Which classification determines whether mathematical ratios are meaningful?