Data analytics is a repeatable way to turn a question into evidence that supports a decision. It includes understanding the problem, checking the data, choosing a suitable method, interpreting results and communicating both findings and uncertainty.
It is not simply making a chart or running a query. A technically correct calculation can still be unhelpful if it answers the wrong question or relies on incomplete data.
1. Start with a question
Imagine an operations manager asks: “Why did we complete fewer orders this week?” A useful analyst clarifies the time period, what counts as a completed order, whether the target changed, and what decision the result will support.
Then define what evidence is needed: daily order volume, staffing, order complexity, processing time and any known system interruptions. These are possible explanatory variables, not proof of a cause.
2. A small synthetic example
The following values are fictional and created only for practice.
| Week | Target orders | Actual orders | Variance |
|---|---|---|---|
| Week 1 | 2,000 | 1,940 | -60 |
| Week 2 | 2,000 | 1,860 | -140 |
| Week 3 | 2,100 | 1,990 | -110 |
Absolute variance is actual minus target. In Week 2 the variance is 1,860 - 2,000 = -140 orders. Relative variance is -140 / 2,000 × 100 = -7%. That identifies a gap, but doesn’t explain its cause.
3. An analysis workflow
- Define the decision and metric.
- Check the dataset for missing values, duplicates, and period mismatches.
- Calculate a suitable comparison.
- Segment the result to locate where it occurs.
- Test hypotheses against evidence.
- Explain the finding, limitation and next action.
4. Practice exercise
A team targets 1,500 completed orders and records 1,425.
- Calculate the absolute variance.
- Calculate the relative variance.
- Name one useful follow-up question.
Answer: The absolute variance is -75 orders; relative variance is -5%. Next, segment by day, team or order type to learn where the gap is concentrated. The data must support any explanation.
Takeaways
Good analytics connects a well-framed question with reliable data, a suitable calculation and a decision. Be explicit about the difference between an observed gap and a hypothesis about why it happened.