Data Types, Sources, and Structure
Lesson 4: Data Types, Sources, and Structure
What kinds of data do analysts work with?
Type | Description | Examples | Common tools |
| Structured | Consistent rows and columns | Spreadsheets, database tables | Excel, SQL |
| Semi-structured | Some organisation, but fields can vary between records | JSON events, XML, log files | Python, SQL with JSON functions |
| Unstructured | No predefined format | Free text, images, audio, PDFs | NLP tools, specialised libraries |
Most entry-level analytics work uses structured data.
Common field types
Field type | Description | Example | Can you calculate with it? |
| Numeric (continuous) | Measured quantity | Revenue, weight, time | Yes |
| Numeric (discrete) | Counted quantity | Number of items | Yes |
| Categorical (nominal) | Labels with no order | Region, product category | Count only |
| Categorical (ordinal) | Labels with a natural order | Low / Medium / High; ratings 1–5 | Count, rank, sometimes median |
| Date / time | Points in time | Order date, timestamp | Yes, for durations and trends |
| Boolean | True / false | Is_returning_customer | Count, percentage |
| Identifier | Uniquely labels a record | Order ID, customer ID | No |
- Numbers may represent measurements (revenue) or labels (a product code).
- A code like 00127 should not be treated as a quantity just because it contains digits. Storing it as a number would also drop the leading zeros.
Where does data come from?
Source | What it contains | Example |
| Transaction systems | Orders, payments, bookings | E-commerce database |
| CRM tools | Customer contacts, leads, deals | Salesforce, HubSpot |
| Web / app analytics | Page views, events, sessions | Google Analytics, Mixpanel |
| Surveys | Opinions and self-reported data | Customer satisfaction forms |
| Spreadsheets | Manual or exported data | Budget trackers |
| Application logs | System events and errors | Server logs |
| Public datasets | Open government or research data | Census data, Kaggle datasets |
| APIs | Data fetched from other services | Weather, exchange rates |
Before using any source, check:
- Owner – who maintains it and who can answer questions?
- Update frequency – real-time, daily, monthly?
- Definitions – what does each field mean?
- Permissions – are you allowed to use it for this purpose?
- Known gaps – missing periods, tracking changes, known errors?
Example
An order table:
order_id | order_date | customer_id | product_category | quantity | unit_price | region |
| 10001 | 2026-07-01 | C204 | Books | 2 | 350 | South |
| 10002 | 2026-07-01 | C118 | Electronics | 1 | 4,999 | North |
Column | Role |
| order_id | Identifies a record |
| order_date | Supports time-based analysis |
| customer_id | Links to the customers table |
| product_category, region | Categories for grouping |
| quantity, unit_price | Support calculations (revenue = quantity × unit_price) |
Data quality questions
Dimension | Question to ask | Example problem |
| Completeness | Are required values present? | 15% of rows have no region |
| Validity | Do values follow the rules? | Negative quantities |
| Consistency | Are values recorded the same way? | "Chennai", "chennai", "CHN" |
| Uniqueness | Are records duplicated? | The same order appears twice |
| Timeliness | Is the data current enough? | Data last refreshed three weeks ago |
| Accuracy | Does it match reality? | Prices differ from the product catalogue |
A date stored as text, a duplicated order, or inconsistent region names can all distort the result.
Key takeaways
- Understand what each column means before calculating anything.
- A field's meaning matters more than its appearance or storage format.
- Always check quality across completeness, validity, consistency, uniqueness, timeliness, and accuracy.
Data Analytics Tutorial
Jebasta








