Data Analysis Basics: Types, Tools & Real-World Impact

What is Data Analysis?

Data analysis is the process of inspecting, cleaning, transforming, and modeling data to discover useful information, draw conclusions, and support decision-making. It bridges the gap between raw data and actionable insights.

Key Components of Data Analysis

1. Data: Raw facts and figures (numbers, text, images, etc.).
2. Tools: Software like Excel, Python (Pandas), R, or SQL.
3. Techniques: Statistical methods, machine learning, visualization.
4. Outcome: Reports, dashboards, or automated decisions.

Types of Data Analysis

Data analysis is classified based on purpose, complexity, and output. Here’s a comparison table:

The Data Analysis Lifecycle

Every analysis follows a structured cycle of steps. Visualize it below:

flowchart LR
    A["1. Define Problem"] --> B["2. Collect Data"]
    B --> C["3. Clean & Preprocess"]
    C --> D["4. Explore & Visualize"]
    D --> E["5. Model & Analyze"]
    E --> F["6. Interpret Results"]
    F --> G["7. Communicate Insights"]
    G -->|"Feedback Loop"| A

Step-by-Step Breakdown

1. Define the Problem

   • Example: A bank wants to reduce loan defaults. Question: Which customer segments are most likely to default?
   • Tool: Problem statement template (e.g., "Reduce churn by 20% in 6 months").

2. Collect Data

   • Sources:
     • Primary: Surveys (e.g., NEPSE investor sentiment polls).
     • Secondary: Public datasets (e.g., World Bank GDP data), APIs (e.g., Daraz sales data).
   • Real Example: Pathao collects rider location, trip duration, and payment data via its app.

3. Clean and Preprocess Data

   • Common Issues:
     • Missing values (e.g., 10% of Ncell customer records have no call duration).
     • Duplicates (e.g., same transaction ID in eSewa data).
     • Inconsistent formats (e.g., "2023-12-31" vs. "31/12/2023").
   • Tool: Python’s pandas or Excel’s "Find & Replace."

4. Explore and Visualize

   • Example: Analyzing Khalti’s transaction failures by hour.
     • Visual: A bar chart showing failure rates peaks at 3–5 PM (likely due to high volume).
     • Tool: Tableau or Python’s matplotlib.

5. Model and Analyze

   • Apply statistical or machine-learning models.
   • Example: Predicting NEPSE stock prices using past trends (time-series forecasting).

6. Interpret Results

   • Example: If a model shows that customers with <3 months tenure churn 3x more, the bank might offer them a loyalty discount.

7. Communicate Insights

   • Tools: Dashboards (Power BI), reports (Word/PDF), or presentations (Canva).

In the Real World

1. eSewa’s Fraud Detection

   • Idea: Anomaly Detection (a type of diagnostic analysis).
   • How: eSewa uses algorithms to flag unusual transaction patterns (e.g., a single user making 50 payments in 1 minute). This reduces fraud by 40% (as per their 2023 report).

2. Daraz’s Supply Chain Optimization

   • Idea: Predictive Analytics.
   • How: Daraz’s AI predicts demand for products like Diwali gifts 3 months in advance, ensuring warehouses stock enough inventory. This reduced out-of-stock items by 25% in 2022.

3. Ncell’s Network Performance

   • Idea: Descriptive + Prescriptive Analysis.
   • How: Ncell analyzes call-drop rates by tower location (descriptive). Then, it uses linear programming to optimize tower placements, reducing drops by 15% in Kathmandu.

Worked Example: Analyzing Ncell Customer Churn

Problem: Ncell wants to identify which customers are likely to switch to NTC. We’ll use descriptive and diagnostic analysis on a sample dataset.

Step 1: Data Collection

Assume we have a dataset with:

• Customer ID, tenure (months), monthly calls, data usage (GB), complaints filed, and churn status (Yes/No).

Step 2: Data Cleaning

• Check for missing values: None here.
• Convert "Churned?" to binary (Yes=1, No=0).

Step 3: Exploratory Analysis

Visual 1: Average calls and data usage by churn status.

graph LR
    A["Churned=No"] --> B["Avg Calls: 250"] --> C["Avg Data: 7 GB"]
    D["Churned=Yes"] --> E["Avg Calls: 160"] --> F["Avg Data: 3 GB"]

Observation: Churned customers use fewer calls and data.

Visual 2: Complaints vs. Tenure (scatter plot).

Insight: Customers with <6 months tenure and >2 complaints are high-risk.

Step 4: Modeling (Diagnostic)

• Hypothesis: Customers with low tenure and high complaints are more likely to churn.
• Test: Use a decision tree to classify churn risk.
• Action: Ncell could target customers in the "High" group with retention offers.

Tools for Data Analysis

Advantages and Limitations of Data Analysis

Exam Tip

1. Define Clearly: Always start with a problem statement (e.g., "Analyze why Pathao drivers in Pokhara have low earnings").
2. Show Steps Visually: Use flowcharts (like the lifecycle above) or tables to organize answers.
3. Link to Nepal: Examiners love real-world ties. Mention eSewa, Daraz, or Ncell in your examples.
4. Avoid Jargon: Explain terms like "EDA" (Exploratory Data Analysis) in simple words.
5. Practical > Theoretical: Focus on how to apply concepts (e.g., "Use a pivot table in Excel to summarize NEPSE data") rather than just definitions.

Key Takeaways

• Data analysis turns raw data into actionable insights for businesses and governments.
• The lifecycle (problem → data → model → action) is universal, from eSewa to Google.
• Visuals (charts, tables, trees) make complex data understandable—always include them in exams.
• Tools like Excel and Python are essential; know when to use each.
• Real-world applications (e.g., Daraz’s demand forecasting) show how theory works in practice.