Skip to content

Latest commit

 

History

History

workshop

Cognite logo

Introduction to Open Industrial Data

Took a course on python programming in first year? Machine learning guru? Evangelical pythonista? Information designer? IoT geek? Welcome, we've got a spot for you all. We need your help to analyze and visualize the industrial world!

Speak up, ask questions and work together. We can't wait to see what you come up with!

Cogniters

Instructions

Although this material references data science topics, we have designed it to cover multiple levels of experience. We start together, but then break out into different specializations, a lot like we do in our teams today at Cognite!

Part 1: Get set up

Part 2: Explore the 1st Stage Gas Export Compressor train on the Valhall platform

System overview We do infographics slightly differently here at Cognite: with real time data!

  • Look through the infographics to find your way to the System Overview
  • Pick a subsystem that you would like to analyze further (labelled on the diagram, e.g. '23-VG-9101') and write down the tag name.

Part 3: Asset data dive

Let's get stuck into the data for our chosen asset. The Cognite Python SDK gives us full access to the industrial data. The notebook "Part 3 - Asset data dive.ipynb" walks us through the basics of how industrial data is represented in CDP.

If you have a notebook environment set up on your laptop, simply clone down this repo and get to work. Otherwise, we recommend Google Colab. Simply reference the url of the notebook in github from colab, or open the notebook directly.

  • Run the notebook from start to finish, exploring the asset hierarchy and timeseries for the asset you chose in Part 2 (use <shift+enter> to step through the cells and run the code snippets).

Part 4: Specialization

Time to split into specializations! Either build an engaging visualization in Operational Intelligence, or continue writing code!

Part 4a: Real time data visualization

In your notebook from part 3, you generated a list of timeseries associated with your asset of choice. Let's head back over to Operational Intelligence and liberate that data by creating an engaging visualization!

  • Research your asset of choice (ABB Oil & Gas Production Handbook)
  • Draw a simplified representation or find a diagram online
  • Create an infographic in OpInt. Be sure to select "Aker BP" when adding time series (tags).
  • Using the timeseries descriptions you loaded in df_asset_children_timeseries in your notebook from Part 3, identify and add time series to your infographic in the correct locations.
  • Publish your dashboard to the world!

Part 4b: Data science

A bit excited about that time series data you saw in Part 3? We get it. Here are a couple ideas to get you started:

  • Data quality investigation
  • Graph analytics
  • Supervised sensor prediction
  • Unsupervised anomaly detection

We have a notebook to get you started.

When you're done

Thanks for participating! Share your work with the world and contribute to the vision of Open Industrial Data!

And don't forget that, like most other companies, we look at Github profiles when we recruit for the Data Science, and strongly encourage you to build up a history of unique work. It could be your foot in the door to your dream job :)