A Browser-Based Application for Engineers to Find Insights in Time Series Data

Written by Michael Risse on July 25, 2016

We have had an increasing number of requests for a simple, quick way to understand Seeq. These come from engineers who want to share Seeq with colleagues or new customers that are trying to figure out where Seeq fits. We hear “is it like this” or “is it like that,” as if Seeq fits into a known software category – a spreadsheet, a programming language, a trending tool, a visualization package, a statistics program.

In fact, Seeq is a little bit of all of these in a new form factor – a browser-based application for engineers to find insights in time series data – and there isn’t a category of existing products where it fits. Instead, Seeq is the first of a new generation of solutions for working with time series data. To help explain that, we’ve created this infographic to explain a typical workflow for an engineer using Seeq. The image depicts 3 work steps but for the purposes of explanation we’ll spread it out to 5.

Data Investigation Using Seeq

First: The Need. We like to celebrate the “aha moments” that Seeq enables, the triumphant discovery at the end of an investigation into root cause or optimization or correlation, and we have those use cases to be sure. But far more often the day of an engineer is dominated by either the urgent: alarms and alerts, or the mundane: creating reports. Daily reports, weekly reports, monthly reports. Investigations, alarms and reports, any of these scenarios can be the impetus to begin work with Seeq, so it begins…

Second: Data Wrangling. “Data wrangling” is borrowed from IT departments and refers to the time spent getting, cleansing, correlating, and structuring data to enable analysis. Our customers say that it is typical to spend 80% of project time on data wrangling, and with Seeq it can happen in just minutes. Just connect to the data sources, of any type, and get the required sensor data from the historian. And statistics from a batch, LIMS, or MES system. And stranded data in a CSV file. And work orders from SQLServer. And data from a cloud IoT system. Seeq makes it easy and fast to assemble and contextualize data, saving hours and days of engineer effort.

Third: The Investigation. With the data assembled, cleansed, and organized, it’s time to use Seeq's investigation features to work with the data. These features include search capabilities that employ advanced algorithms to find pattern matches, limit conditions, and define time periods of interest. There is a scripting language to write equations to combine and analyze results and scatter plots to visualize data sets. And all of this can be shared with colleagues, in real time, through Seeq’s brower-based experience, so access to sharing insights and expertise is literally a mouse click away.

Fourth: Do Something About it. At the conclusion of the work, the next step is to do something with what has been discovered. The most likely choices are either to monitor for a situation that’s been identified (so if it occurs again the engineer is notified) or to document the effort for access and reuse by others. Seeq supports both of these choices. With the help of Seeq Analytics Engineers, you can set up a monitoring solution to track a particular asset or sensor condition in the context of other data sets. Or you can save your work in Seeq, capturing the history of your investigation, including annotations that your colleagues can search, workbooks with names to identify their contents (Valves, Minnesota Plant), or custom equations and functions they can re-use in the future.

Fifth: Put the Insights Where They Need to be. Not every user that benefits from the insights that Seeq enables is an engineering user, or there may be a last step in the workflow that requires a particular calculation or visualization that works in Microsoft Excel or another product. Which is no problem because Seeq can export data to Excel worksheets or share data with Microsoft PowerBI or Tableau through an OData interface. And, in response to customer requests, Seeq now supports publishing of Seeq insights into OSIsoft Coresight for dashboard distribution.

Using Seeq for the arc of the engineer's work process from impetus to sharing or publishing a report or insight results in hours and hours of saved time. From 5 hours to 30 minutes in Excel, from 3 days to 15 minutes in report writing. On the insight side it’s harder to calculate the time saved in minutes because frequently the issues being investigated were considered so hard or challenging that our customers hadn’t yet even attempted the work. This is the case with the Avangrid case studythe question on the value of curtailment periods had lingered for a year before Seeq was used to find the answer in just an hour. We hear these recurring stories from our customers: problems previously considered too hard to solve aren’t anymore, and hours and days melt into minutes with Seeq.