Data Analysis Fundamentals
What Data Analysis Actually Is and Why It Matters
Data analysis is the process of examining, cleaning, transforming, and interpreting data to discover useful information, draw conclusions, and support decision-making.
It's not magic. It's not just for experts. It's a learnable skill.
The Purpose of Analysis
Answer Questions
"Why did sales drop last month?" "Which marketing channel is most effective?" "What predicts customer churn?"
Analysis turns vague questions into specific answers.
Find Patterns
Data contains patterns humans can't see by scanning rows. Analysis reveals them.
Support Decisions
Gut feeling vs. data-informed decision. The latter is usually better.
Measure Reality
What's actually happening vs. what you think is happening? Data tells you.
Types of Data Analysis
Descriptive Analysis
Question: What happened?
Methods: Summarizing, counting, averaging, visualizing historical data.
Example: "We sold 10,000 units last quarter, up 15% from the previous quarter."
Diagnostic Analysis
Question: Why did it happen?
Methods: Comparing segments, drilling down, finding correlations.
Example: "Sales increased because our email campaign drove a 40% increase in website traffic."
Predictive Analysis
Question: What will happen?
Methods: Forecasting, trend projection, statistical modeling.
Example: "Based on current trends, we project 12,000 units next quarter."
Prescriptive Analysis
Question: What should we do?
Methods: Scenario modeling, optimization, recommendations.
Example: "Increasing email frequency by 20% while maintaining current content quality would maximize conversions."
Most analysis starts descriptive, becomes diagnostic, and occasionally extends to predictive.
Key Concepts
Variables
A variable is anything that can be measured or categorized.
Quantitative variables: Numbers (revenue, age, quantity) Categorical variables: Categories (region, product type, gender)
Data Points (Observations)
Each row in your data — a transaction, a customer, a day.
Metrics vs. Dimensions
Metrics: What you measure (sales, count, average) Dimensions: How you slice it (by region, by month, by product)
Populations vs. Samples
Population: All possible data (every customer ever) Sample: A subset (customers in your database, survey respondents)
Most analysis works with samples. This matters for how confident you can be.
The Analysis Process
1. Define the Question
What are you trying to learn? Be specific.
Vague: "Tell me about our customers." Specific: "Which customer segments have the highest lifetime value?"
2. Gather Data
What data do you need? Where is it? How will you get it?
3. Clean and Prepare
Real data is messy. You'll spend significant time cleaning.
4. Explore
Look at the data. Summarize. Visualize. Notice patterns and oddities.
5. Analyze
Apply appropriate methods to answer your question.
6. Interpret
What do the results mean? What conclusions can you draw? What are the limitations?
7. Communicate
Share findings in a way others can understand and act on.
Common Mistakes
Starting Without a Question
Fishing in data rarely finds anything useful. Start with a question.
Ignoring Data Quality
Garbage in, garbage out. Bad data produces bad analysis.
Confusing Correlation with Causation
Just because two things move together doesn't mean one causes the other.
Overfitting the Data
Finding patterns that exist only in your specific data, not in reality.
Confirmation Bias
Finding what you expect to find because you look for it.
Ignoring Context
Numbers without context are meaningless or misleading.
AI Prompt: Analysis Orientation
Help me plan a data analysis project.
My question: [What you want to learn]
Data I have: [Describe your dataset]
Context: [Relevant background]
How results will be used: [Decision or action this supports]
Help me:
1. Clarify and refine my question
2. Identify what the data can and can't answer
3. Suggest an analysis approach
4. Flag potential issues or limitations
5. Outline next steps
What's Next
Analysis requires a particular way of thinking. Let's develop it.
Next chapter: Thinking like an analyst.