Statistics Basics: Data Types, Scales & Collection

What is Statistics?

Statistics is the science of collecting, analyzing, interpreting, and presenting data to make informed decisions. It helps businesses understand trends, risks, and opportunities.

Types of Statistics

mindmap
  root((Statistics))
    Descriptive
      "Summarizes data (mean, median, charts)"
    Inferential
      "Makes predictions (hypothesis testing, sampling)"

Visual:

Levels of Data Measurement

Data can be classified into four scales, each with different analysis capabilities:

Worked Example 1: Classify the following data types:

1. Blood type (A, B, AB, O) → Nominal (no order, categories only).
2. Movie ratings (1★ to 5★) → Ordinal (ordered but no fixed difference between ratings).
3. Shoe sizes (36, 37, 38) → Interval (ordered, equal intervals, but no true zero).
4. Company profits (₹100,000, ₹200,000) → Ratio (true zero, meaningful ratios).

Visual:

Methods of Data Collection

Businesses gather data through three primary methods:

1. Surveys (Questionnaires)

• Pros: Cost-effective, large sample size, flexible.
• Cons: Response bias, low response rate.
• Example: eSewa collects user feedback via app surveys to improve service.

2. Experiments

• Pros: Controlled environment, causal relationships.
• Cons: Expensive, time-consuming.
• Example: Nepal Rastra Bank tests new loan policies on a small group before nationwide rollout.

3. Observation

• Pros: Natural behavior, no interference.
• Cons: Time-intensive, subjective.
• Example: Pathao observes driver routes to optimize delivery times.

Comparison Table:

In the Real World

1. eSewa (Data Scales & Surveys):

   • Uses ratio data (transaction amounts in ₹) to analyze spending patterns.
   • Conducts surveys to classify user satisfaction (ordinal data: Poor/Fair/Good).

2. Khalti (Interval Data & Experiments):

   • Tracks transaction times (interval data) to improve payment processing.
   • Runs A/B tests (experiments) to compare new vs. old UI designs.

3. Nepal Stock Exchange (NEPSE) (Ratio Data & Observation):

   • Analyzes stock prices (ratio data) to predict market trends.
   • Observes trading volumes to identify seasonal patterns.

Worked Example 2 (Real-World Tie): NTC wants to analyze internet speeds across Nepal.

• Data Type: Ratio (speeds in Mbps have a true zero and allow ratios).
• Collection Method: Experiments (controlled tests in Kathmandu, Pokhara, and rural areas).
• Analysis: Compare mean speeds using descriptive statistics to identify slow regions.

Visual (NTC Speed Data):

Key Formulas and Definitions

1. Population vs. Sample:

   • Population: Entire group (e.g., all TU students).
   • Sample: Subset of population (e.g., 500 TU students surveyed).

2. Variables:

   • Dependent Variable (Y): Outcome (e.g., sales revenue).
   • Independent Variable (X): Input (e.g., advertising spend).

Visual (Variables in Business):

flowchart TD
  A["Independent Variable\n(X: Advertising Spend)"] --> B["Dependent Variable\n(Y: Sales Revenue)"]
  B --> C["Goal: Find relationship\nbetween X and Y"]

Common Pitfalls

• Misclassifying data scales: Using mean on nominal data (e.g., averaging blood types) is invalid.
• Bias in surveys: Leading questions (e.g., "Don’t you hate slow eSewa payments?") skew results.
• Small sample size: A survey of 10 people cannot represent all Nepalese internet users.

Exam Tip

1. Memorize the four data scales and their allowed operations—this is a high-weightage topic in TU exams.
2. Practice classifying real-world data (e.g., "Is age ratio or interval?").
3. Expect questions on:
   • Differences between descriptive vs. inferential statistics.
   • When to use surveys vs. experiments (e.g., "How would Ncell test a new tariff plan?").
4. Diagrams matter: Draw Venn diagrams for data scales or flowcharts for data collection methods in exams.

Summary Checklist: ✅ Define statistics and its two types. ✅ Classify data into nominal, ordinal, interval, ratio. ✅ Explain surveys, experiments, observation with business examples. ✅ Avoid common mistakes (e.g., using mean on nominal data). ✅ Link concepts to Nepali companies (eSewa, NTC, NEPSE).