We are hiring! See our vacancies
3 min. read
Noortje Vollenberg
Last update: 19-03-2021

The 4 types of information you can extract from machine data

Have you started data logging, or are you thinking about setting up a data strategy? One of the most common worries people experience during this process is what to do with the machine data after it’s collected.

What is machine data?

"Machine data is data which is generated by machines without any human involvement. The most common types of data come from sensors or meters, interactions by machine or user input (e.g. switches), algorithms calculated using other data, and data from automation tasks.”. This is how we described machine data in a previously post.

Machine data can give insight into its status, performance, capacity, production or parameters like temperature, pressure or vibrations. The collective data from these sensors, switches, algorithms and automation can be stored and visualised using an Industrial IoT platform such as IXON Cloud.

But what’s next?

Don’t let your valuable machine data gather dust

Your raw data is basically just a compilation of facts. Without interpretation, the data has zero meaning and loses all its value. To turn your data into valuable information, machine data analysis is required.

This term, data analysis, can spike apprehension – or even fear – in some people. Because let’s be honest, if you don’t know exactly what you’re getting into – you will likely not get into it at all.

Except that not getting on the data analysis bandwagon could have an immense, negative impact on your future business. There’s so much value to be extracted from machine data.

You just need to know what you’re doing.

The basics of getting started with machine data analysis

If you’re extracting these four types of information from your data, you’ve completed a proper and highly valuable data analysis:

  • What happened?
  • Why did it happen?
  • What will happen?
  • What needs to happen?

Don’t worry about getting to all four right away. Start small and easy, and scale up as you go!

1) Get a clear picture of the situation using descriptive information

The first step is to take a look at what happened. In other words, get descriptive information from the machine’s data. This provides you with basic facts and a description of events. What happened? When and where did it happen? Who was there?

You can find such information by analysing data like an oven’s temperature, and its increase of 10 degrees from one hour to the next yesterday evening. Or the status of a certain conveyor belt in the machine, which stopped in the middle of production.

Once you’re logging such data tags and visualising them in condition monitoring dashboards, it’s easy to extract these facts.

machine data analytics machine data analytics

2) Dig down to the root cause with diagnostics

This requires a lot of persistence and asking the same question over and over again: “why?”. To get down to the root of the issue, you will have to find answers to what happened and why it happened – diagnostic information.

To stick to one of our previous examples: why did the machine stop in the middle of production?

The first step is to figure out which factors can influence this outcome. Second, you look at the data connected to these factors to figure out their status leading up to and during the breakdown.

Once you’ve interpreted this data and learned new information, you go back to step one. Apply the same steps to this factor and dig down another level. Eventually, you will end up with the root cause of the event.

3) Predictive information about future machine events

If nothing changes, and everything keeps running the same way, what will happen? Predictive information describes what’s going to happen, when it’s going to happen and why.

Obtaining this information is more difficult, but significantly adds value to your machine data.

Find patterns in historical data to predict your machine’s future actions. Once you figured out a pattern, it’s easily applied to predict future events. Follow the standard timeline and you’ll figure out when you can expect the event to occur again.

Live Machine Condition Monitoring Live Machine Condition Monitoring

4) Stay ahead of future issues with prescriptive insights

Ready to step it up a notch?

The fourth, and most difficult, type of information you can extract from machine data is prescriptive information. This means asking what needs to happen next to stop a bad event from occurring, or to ensure a good event occurs.

You won’t regret analysing your data for predictive insights now. With this information at hand, you can optimise processes to ensure that you step in before an event occurs. If you’ve found gaps, data that indicates where you can improve the machine, it’s even possible to optimise your future machines.

Less downtime and better machines – what more could you ask for?

Get started with machine data logging and visualisation on IXON Cloud

Start extracting and visualising your machine data for in-depth analysis today using the IIoT platform for industrial machines. Explore our pre-build dashboards from machine data in our free product tour. 

((Start a Free Product Tour))

Tip! Try out the new Virtual Demo Device to truly get the whole experience.

Related customer stories of IXON's Data Logging and Dashboard feature