Types of Analytics

Today, we’ll discuss types of analytics and their importance in turning raw data into actionable insights.

This diagram shows four types of analytics, ranked by their difficulty level and the value they provide. Let’s go through them step by step.

1. Descriptive Analytics: What Happened?

At the base of the analytics hierarchy is descriptive analytics. This is the simplest form of analytics and helps us understand what happened by interpreting historical data.

• Purpose: To summarise past events and identify patterns.
• Example: A smart thermostat that shows last week’s energy usage patterns.
• Methods: Charts, graphs, and dashboards that clearly show past performance.

This type of analytics is great for reviewing the past, but it doesn’t tell us why something happened or what will happen next.

2. Diagnostic Analytics: Why Did It Happen?

Moving up, we have diagnostic analytics, which looks at why something happened. It’s more complex than descriptive analytics because it requires diving deeper into the data.

• Purpose: To discover relationships and identify the causes behind past events.
• Example: Analysing why a specific day’s energy usage was higher than average by correlating data with external factors like weather.
• Methods: Data discovery, drill-down techniques, and correlation analysis.

This stage helps us make sense of the past by understanding the root causes of trends and anomalies.

3. Predictive Analytics: What Will Happen?

Next is predictive analytics, which focuses on forecasting future outcomes. This is where analytics becomes proactive rather than reactive.

• Purpose: To predict what might happen based on current and historical data.
• Example: A smart thermostat forecasting energy usage for the upcoming week based on weather patterns and past behaviour.
• Methods: Statistical modelling and simulations.

By identifying trends and patterns, predictive analytics helps us make informed predictions.

4. Prescriptive Analytics: How Can We Make It Happen?

At the top is prescriptive analytics, the most advanced type. This involves predicting outcomes and recommending actions to achieve desired results.

• Purpose: To decide the best course of action based on predictions.
• Example: A smart thermostat automatically adjusting settings to save energy while maintaining comfort.
• Methods: Machine learning and AI to analyse probabilities and make decisions.

Prescriptive analytics provides the highest value by enabling automated and data-driven decisions.

IoT and Analytics

This diagram also highlights how analytics works in an IoT platform:

1. Sensors: Collect data from various sources like temperature, humidity, or movement.
2. IoT Platform: Acts as a central hub to process and store the data.
3. Analytics Engine: Applies these four types of analytics to generate insights and drive decisions.

Final Thoughts

Each type of analytics builds on the previous one, moving from simple data interpretation to actionable decisions. The value increases as we move up the hierarchy, as does the complexity.

Question to consider: Which type of analytics is most valuable in your industry, and how can you implement it effectively? Let’s discuss it!

[Note: Download IoT Notes by Mazlan Abbas ]