The Internet of Your Things: From Oil Pipelines to Data Pipelines
The Microsoft Global Energy Forum 2015 was held on Wed April 22, 2015 in Houston, TX. This conference hosted over 400 business and technology leaders in the oil & gas domain. I was honored to have written code and presented at the opening keynote demo. In this post, I will provide resources for the technologies discussed in our demo.
First, some background. I got involved from my work building Pipe Dreams at a Microsoft hackathon. From Jan – April 2015, I’ve been working with a team of people across Microsoft to implement an IoT/Azure scenario of monitoring and applying predictive analytics using Azure Machine Learning for maintenance of an oil/gas pipeline. The team included people from DX, sales, consulting, UX/design, the Microsoft Technology Centers, Americas Office of the CTO, WW Industry, Gulf Coast District Industry (oil & gas), and District STU. We met on evenings and weekends, to build out an end-to-end IoT demonstration of an oil & gas pipeline with gorgeous website frontend, Azure machine learning, Dynamics xRM, PowerBI, Skype translation, Microsoft Project, O365, and more. Data flowed through Azure from Event Hubs to Stream Analytics to an Azure SQL Database, where it was processed by Azure machine learning and displayed by the web frontend. In addition, we showcased some of our Office functionality with PowerBI analytics and Skype translation to allow a Spanish-speaking employee and English-speaking employee to collaborate.
After months of effort, we presented this work as the main keynote demo immediately following and supporting Joseph Sirosh (CVP of Azure Machine Learning) at the Microsoft Global Energy Forum in Houston, TX. We performed a 20-minute scenario that centered on a new employee starting work monitoring an oil/gas pipeline, and the rest of the engineers explaining the integrated technical solution built on the Microsoft platform to enable pipeline monitoring.
NOTE: More resources coming! The source code for the website and the machine learning model will be released by September 2015.
Skype Translation Video from rehearsal: http://blogs.msdn.com/b/jennifer/archive/2015/05/13/skype-translator-in-action.aspx
During the keynote demo, I also showed:
Other articles that I didn’t show during the keynote, but may be useful:
- "Get started with Azure Machine Learning" article
- Predictive maintenance model from Azure Machine Learning gallery
- Blog post on anomaly detection using machine learning to detect abnormalities in time series data