Types of Data Analytics

What are the main types?

Four common categories are descriptive, diagnostic, predictive, and prescriptive analytics.

• They classify the question being asked, not the tool being used.
• They increase in complexity and, usually, in value to decision-making.
• A single project often uses more than one category.

Descriptive analytics: What happened?

Summarises past activity.

• Examples: monthly revenue, average delivery time, website visits, number of support tickets
• Common methods: totals, averages, counts, percentages, trends over time
• Common outputs: reports, KPI cards, dashboards
• Most reporting and dashboard work is descriptive.

Diagnostic analytics: Why did it happen?

Explores possible reasons for a result.

• Example: If sales fell, compare product categories, regions, prices, stock availability, and traffic sources.
• Common methods: segmentation, drill-downs, comparisons between periods, correlation analysis
• Caution: A pattern can suggest a cause, but it does not automatically prove causation.

Predictive analytics: What might happen?

Uses historical patterns and statistical or machine-learning methods to estimate a future or unknown outcome.

• Examples: forecasting next month's demand; estimating the probability a customer will cancel a subscription
• Common methods: time-series forecasting, regression, classification models
• Caution: Predictions are estimates with uncertainty, not guarantees.

Prescriptive analytics: What should we do?

Compares possible actions using constraints, predictions, and decision rules.

• Example: A delivery team evaluates how to assign drivers to reduce delays while respecting working-hour limits.
• Common methods: optimisation, simulation, scenario analysis, decision rules
• Caution: Recommendations depend on the quality of the predictions and the constraints you set.

Comparison

• The categories build on one another.
• Not every project needs all four. Many valuable analyses are purely descriptive or diagnostic.