Data science is a term heard more and more frequently nowadays. Our team has built out features incorporating data science in our Simplicity Cloud™, with the goal to provide as much insights into your data as possible. In this blog we will discuss what data science is and how to use our data science features to your advantage. To start, let’s define data science in simple terms:
Data science provides insights into the data to derive equipment health. It shows you visually how your equipment is performing compared to past performance, as well as predicting the future. Data science drives preemptive maintenance, provides forecasting, real time predictions, anomaly detections, pattern recognitions, and more.
This dashboard has a lot of data science going on. Its main purpose is to indicate overall asset health at a glance. When looking at the cards at the top of the screen, both the color and the line in each card help indicate how each metric is trending. The green color signifies less change over the specified time period, and the red color signifies more change. A yellow color indicates moderate change.
Each metric card consists of the average metric value, percent change, and a sparkline to signify how the metric is trending over the selected time period. The average value can be used to compare against expected baseline values while the percent change and sparkline can be used to indicate how stable or unstable your equipment is running. The larger chart below shows the sparkline blown up in more detail, including simple linear regression. This is a linear model of how the series trends over time. You can toggle the line on/off, as well as the alarm thresholds on and off in the graph. On the right side you can see recent alerts and notes.
Tip: To be thorough, checking up on equipment that is colored red/green or contains large spikes in the graph is recommended to catch issues before they become problematic.
The change indicator band is pictured here. The thin horizontal line in the horizontal center sets a baseline for comparing the min and max values to their respective distributions. The lines above and below the line indicate the most extreme standardized distances of the min and max values from their respective expected value ranges, across all metrics. The height of the bars indicates how much of an outlier that data point is, and larger spikes assist viewers in determining where to focus diagnostic efforts.
Tip: If you see these lines on your own sensor’s chart, begin with these outliers to start painting a picture of what happened to produce that measurement.
What’s next with data science?
Our data scientists are setting the foundation for machine learning and anomaly detection. Our Simplicity Cloud is constantly evolving and working to provide our customers a user friendly platform to better understand their equipment. We’ve got some awesome new features currently in the works, but under lock and key for now.