Descriptive Statistics

Descriptive statistics is a branch of statistics that summarizes and describes the main features of a dataset.

• It does not involve prediction or inference.
• Focuses on understanding data characteristics through numerical summaries and visualizations.
• It is a key step in Exploratory Data Analysis (EDA).

EDA = Descriptive Statistics + Data Visualization

Dataset and Data Types

A dataset is a collection of data objects (examples: records, patterns, samples, observations).
Each data object is described by attributes (features or properties).

Example: Sample Patient Table (Table 2.2)

Types of Data (Figure 2.1)

Data is broadly classified into:

1. Categorical (Qualitative) Data

• Describes qualities or labels, not measurable.
• Further classified into:
  • Nominal Data: No inherent order.
    e.g., Patient ID, Blood Group, Gender.
    • Only equality comparisons (=, ≠) are valid.
  • Ordinal Data: Has a meaningful order.
    e.g., Fever = {Low, Medium, High}
    • Can be ranked, but exact differences are unknown.

2. Numerical (Quantitative) Data

• Represents measurable quantities.
• Subtypes:
  • Interval Data: Numeric values with meaningful differences, but no true zero.
    e.g., Temperature in Celsius or Fahrenheit.
    • Operations allowed: +, –
  • Ratio Data: Has a meaningful zero and allows all mathematical operations.
    e.g., Age, Height, Weight.

Based on Values

Based on Number of Variables (Figure 2.2)