Industrial Time Series Analysis – Part 1

Most modern enterprise-scale manufacturing processes include the capture of multi-modal sensor data. There is often business value in this data, such as improving yield, optimizing the timing of cleaning cycles, or maximizing throughput.

Industries from every sector have harnessed the power of RapidMiner to perform predictive or prescriptive analytics on time series sensor data and hence gain an advantage in competitive markets.

Part 1: Data Prep, Batching, and Feature Engineering

In Part 1 of this Lightning Demo, Dr. Martin Schmitz (Data Scientist at RapidMiner) will begin with a raw industrial time series data set and show how to use RapidMiner to prep the data for modeling including:

  • Equalizing time stamps
  • Windowing & batching
  • Extracting relevant features such as slopes, peaks, and FFT transformations for modeling