BlogData Analytics BasicsChapter 2
BASICS · CHAPTER 2Beginner

Types of Data in Analytics

Structured vs unstructured, qualitative vs quantitative, discrete vs continuous, and the 4 measurement scales — with real Indian examples and why it matters for analysis.

structured unstructuredqual quantdiscrete continuousmeasurement scalesin practicefaq
DATA ANALYTICS SERIES:← Ch 1: What is DACh 2: Types of Data ←Ch 3: The DA Process →

Structured vs Unstructured Data

The most fundamental distinction in data analytics is between structured and unstructured data. It determines the tools you use, the difficulty of the analysis, and how much preparation the data needs before you can work with it.

Structured Data
Organised in rows and columns
CHARACTERISTICS
  • Fits neatly into a table or spreadsheet
  • Every record has the same fields
  • Easily queried with SQL
  • Stored in databases (MySQL, Oracle, SAP)
INDIAN EXAMPLES
  • Bank transaction records (account no., date, amount, type)
  • E-commerce orders (order ID, product, city, price)
  • Employee payroll (employee ID, salary, department, month)
  • Hospital patient records (patient ID, ward, diagnosis, cost)
Unstructured Data
No fixed format or schema
CHARACTERISTICS
  • Cannot be stored directly in a table
  • Each item may be completely different in length/format
  • Requires NLP, image recognition, or audio processing
  • Stored in data lakes, object storage (S3, Azure Blob)
INDIAN EXAMPLES
  • Customer reviews on Amazon India or Swiggy
  • Call centre recordings from bank helplines
  • WhatsApp complaints sent to customer care
  • Medical CT scan or X-ray images
  • News articles about company performance
SEMI-STRUCTURED DATA

Between the two extremes sits semi-structured data — it has some organisation but does not fit a rigid table. JSON and XML files are examples. A customer API response from Razorpay or Paytm that contains a mix of text fields, nested objects, and arrays is semi-structured. It requires some transformation before it becomes a clean table for analysis.

Qualitative vs Quantitative Data

Within structured data, every column is either quantitative (numeric, measurable) or qualitative (categorical, descriptive). This distinction determines which calculations make sense and which charts to use.

Quantitative (Numeric)Qualitative (Categorical)
DefinitionMeasured in numbersRepresents categories or labels
Can you average it?Yes — average revenue, average delivery daysNo — you cannot average "city name"
Indian examplesOrder amount (₹), delivery time (days), age, quantity, NPS scoreCity name, product category, payment method, gender, sales rep name
Best chartsLine chart (trends), histogram (distribution), scatter plotBar chart (comparison), pie chart (proportion), heatmap
SQL operationsSUM, AVG, MIN, MAX, STDEVCOUNT, GROUP BY, DISTINCT, COUNT DISTINCT
Excel operationsAVERAGE, SUM, STDEV, percentileCOUNTIF, pivot table count, frequency table

Discrete vs Continuous Data

Quantitative data has a further subdivision — discrete (countable, whole numbers) and continuous (any value within a range, including decimals).

Discrete Data

Countable values that cannot be split into smaller meaningful parts. You can have 5 orders or 6 orders, not 5.3 orders.

EXAMPLES
  • Number of orders placed per day
  • Number of employees in a branch
  • Number of customer complaints per week
  • Number of items in a cart
Continuous Data

Can take any value within a range — including fractions and decimals. Measured rather than counted.

EXAMPLES
  • Revenue (₹14,327.50 — decimal values are valid)
  • Temperature in a warehouse (28.4°C)
  • Delivery time in hours (1.7 hours, 3.25 hours)
  • Customer age (can be expressed as 24.5 years)
WHY IT MATTERS: Discrete data is best shown with bar charts. Continuous data is best shown with histograms (which show distribution) or line charts (which show trends). Confusing the two leads to charts that are technically wrong — for example, a line chart connecting discrete daily order counts implies there are values between days, which does not exist.

The 4 Measurement Scales — Nominal, Ordinal, Interval, Ratio

The measurement scale of a variable determines which mathematical operations are valid. This matters both for choosing the right calculation and for avoiding misleading results.

Level 1Nominal Scale

Categories with no order or ranking. The values are just labels.

✓ YOU CAN
Count how many fall in each category
Find the mode (most frequent category)
Compare frequencies
✗ YOU CANNOT
Rank or order the categories meaningfully
Calculate average or difference
Indian examples: City (Delhi, Mumbai, Bangalore), Payment method (UPI, card, COD), Product category (Electronics, Apparel, FMCG)
Level 2Ordinal Scale

Categories that have a meaningful order — but the gap between ranks is not equal or known.

✓ YOU CAN
Rank and order categories
Find median and mode
Compare which is higher or lower
✗ YOU CANNOT
Calculate average meaningfully (the gap between "Poor" and "Average" may not equal the gap between "Average" and "Good")
Multiply or divide values
Indian examples: Customer satisfaction: Poor / Average / Good / Excellent. Credit rating: A / B / C. Employee performance: Below Expectations / Meets / Exceeds.
Level 3Interval Scale

Numeric scale with equal gaps between values, but no true zero point. Zero does not mean "none of the thing."

✓ YOU CAN
Add and subtract values
Calculate mean, median, standard deviation
Measure the exact difference between two values
✗ YOU CANNOT
Divide or multiply meaningfully (30°C is not "twice as hot" as 15°C)
Say one value is a multiple of another
Indian examples: Temperature in Celsius or Fahrenheit. Year (year 0 does not mean "no time"). IQ score.
Level 4Ratio Scale

Numeric scale with equal gaps AND a true zero point. Zero means "none of the thing."

✓ YOU CAN
All mathematical operations: add, subtract, multiply, divide
Calculate mean, median, SD, percentile
Make ratio comparisons (₹10L is twice ₹5L)
✗ YOU CANNOT
Nothing — this is the most powerful scale
Indian examples: Revenue (₹0 means no revenue). Delivery time in minutes (0 means instant). Age in years. Number of orders. Weight. Distance.

What Data Types Look Like in a Real Indian Dataset

Here is a sample e-commerce orders table — the kind of data you would find in a Flipkart or Myntra internal database. Notice how each column has a different data type, and that data type determines exactly what you can do with that column.

ColumnExample ValueTypeScaleValid Operations
order_idOD12345678QualitativeNominalCount, deduplicate — never average
order_date2026-08-15Structured (date)IntervalDifference between dates, extract month/year, trend over time
cityNoidaQualitativeNominalCount by city, filter, group by — never rank
categoryElectronicsQualitativeNominalCount, GROUP BY, frequency distribution
amount_inr4299.00Quantitative (continuous)RatioSUM, AVG, MAX, MIN, growth %, ratio comparison
quantity2Quantitative (discrete)RatioSUM, AVG — whole numbers only
delivery_days3Quantitative (discrete)RatioAVG, MAX, SLA breach count
rating4 (out of 5)Qualitative/Quantitative (debated)OrdinalMedian, mode — average is technically incorrect but common in practice
payment_methodUPIQualitativeNominalCount by payment type, share %
returnedYes / NoQualitative (binary)NominalReturn rate = returned count / total count
Continue the Series
← Ch 1: What is DACh 3: The DA Process →

Frequently Asked Questions

What is the difference between structured and unstructured data?

Structured data is organised in a predefined format — rows and columns in a database or spreadsheet. Every entry follows the same structure: a sales record has an order ID, date, amount, and city. Unstructured data has no fixed format — customer reviews, WhatsApp messages, call centre recordings, photos, and emails are all unstructured. Structured data is easy to query with SQL and analyse in Excel. Unstructured data requires text analysis, natural language processing, or image recognition tools. About 80% of business data worldwide is unstructured, but most data analyst roles today work primarily with structured data.

What is the difference between qualitative and quantitative data?

Quantitative data is numeric and can be measured — revenue, temperature, delivery time, number of orders. You can add, subtract, average, and compare quantitative data with mathematical operations. Qualitative data is descriptive and categorical — customer feedback, product category, city name, employee department. Qualitative data tells you what type or category something belongs to. In analytics, quantitative data is used for aggregation and trends, qualitative data is used for segmentation and grouping. Most datasets have both — for example, a sales table has quantitative columns (amount, quantity) and qualitative columns (city, category, sales rep name).

What are the 4 measurement scales in data analytics?

The 4 measurement scales are: (1) Nominal — categories with no order (city names, product categories, gender). You can count and compare frequencies but not rank them. (2) Ordinal — categories with a meaningful order but unequal gaps (satisfaction rating: Poor / Average / Good / Excellent). You can rank but not calculate the exact difference between ranks. (3) Interval — numeric scale with equal gaps but no true zero (temperature in Celsius, year). You can add and subtract but not meaningfully divide. (4) Ratio — numeric with equal gaps AND a true zero (revenue, weight, delivery time in minutes). All mathematical operations apply. Most analyst work involves nominal, ordinal, and ratio data.

Why does the type of data matter for analysis?

The type of data determines which analysis methods and visualisations you can use. You cannot calculate the average of a nominal variable like city name — but you can count frequencies and show a bar chart. You can calculate the average of a ratio variable like revenue and show a trend line. Choosing the wrong analysis method for the data type produces misleading results — for example, averaging ordinal satisfaction scores (treating "Poor/Average/Good/Excellent" as 1/2/3/4) is a common mistake that distorts insights. Understanding data types is also essential for knowing how to clean data — a revenue column that contains text values like "NA" or "₹" signs needs different cleaning than a category column with typos.

EVIKA ACADEMY · NOIDA SECTOR 51

Learn analytics from scratch with live guidance

SQL, Python, Power BI, and real projects — structured training for Indian job market. Free demo class available.

Book Free Demo Class →
🎓 Free Demo Class — Online & Offline · Noida Sector 51