IBM has introduced a new tool for data enthusiasts and scientists called “IBM Data Science Experience“. This is a cloud based application that provides multiple open source notebooks namely JUPYTER, ZEPPLIN, RSTUDIO IDE for the scientists to pick and start analyzing historical data in Python, R or Scala. In addition, the scientists can collaborate with colleagues to share and refine their work to glean insights rapidly and accurately.
We have now introduced a recipe where an embedded copy of JUPYTER (open source notebook) can be used to analyze historical data from your devices. The recipe engages a standard Z Score algorithm that identifies thresholds of normal and abnormal behavior of the device. These thresholds are fed into the IBM Watson IoT Platform as rules that analyze real time events and generate alerts (emails or trigger workflows) where devices show deviation in behaviors. I encourage the use of IBM Data Science Experience to detect time series anomalies for your IoT missions and am eager to hear your feedback.