Trend Detection on Stock Prices

May 29, 2017

Squirro’s trend detection engine was originally developed for anomaly and trend detection in operational data such as our Service Insights proposition applied to ITSM environments, like ServiceNow.

Here’s a try to apply the engine to time series data normally referred to as stock price data. In this example we took the last 7 years of Salesforce stock price data.

We run both a prediction from today May 2017 onwards and one to compare what the engine predicted a year ago (May 2016) for today. Here’s the result for the May 2016 prediction: The engine predicted the price for May 30 2017 to be 84.63 with a maximum threshold of 98.31. The stock price per last Friday’s (May 26, 2017) close was 90.83…

But check yourself:

Dashboard 1 – Salesforce Share Trading 

Dashboard 2 – Price Trend Comparison