ESB Data Science team win Microsoft Open Hack with TCS Partners

Problem solving and finding solutions is at the heart of what ESB does. From the construction of Ireland’s first national hydro-electric station in 1925 to predicting output from turbines, we thrive on exploring and solving customer-focused solutions.

Hackathons are now viewed as the best creative environment to help innovators discover ‘the next big thing.’ Taking place over a two or three day period, teams are tasked with creating a solution to specific challenge and have to pitch their idea to a panel of experts.

The Microsoft Hackathon

To test the mantle of ESB’s Data Science team, they competed in the first Microsoft Machine Learning Open Hack in Ireland. This hack event focused on solving problems in the computer vision space, allowing like-minded developers to explore such work as Data Wrangling, Cognitive Services, Azure Machine Learning Workbench, Custom Machine Learning and Neutral Network Solutions on machine learning frameworks.

The hackathon was broken into seven challenges, each designed to increase in difficulty as the team progressed through the competition over the three days.

  

ESB’s Data Analytics Team Solution

Much of the work undertaken by ESB’s Data Analytics team is focused on completing Data Science use cases in the areas of Asset Management, Trading and HR. The team’s background and experience in this area was put to good use during the Microsoft Hack, as was their knowledge and awareness of all things safety related.

The team consisted of ESB staff and our Data Analytics Partners  (Tata Consultancy Services) and comprised of Dibs Bose, Quentin Bragard, Kevin Quinlan, Aine Matthews, and James Conway. The team's tasks included creating and deploying a multi-class classifier for mountaineering equipment as part of their first challenge.

The second task focused on image tagging where the team had to create a model to tag safety helmets on images of climbers. After pre-processing the images using padding and colour balancing techniques, the team used the Microsoft Visual Object Tagging Tool to create a large dataset. They applied a state-of-the-art network in object detection, Faster R-CNN, on our dataset and obtained the best accuracy in the challenge.

Data Drives Company Performance

Data science is transforming the way we live and hence why more people are embracing this new career field. From cool functions, like self-driving cars to improving safety by better monitoring, the opportunities are endless in this area while they are an integral part of any company.

Check out https://www.esb.ie/careers for more information on ESB and opportunities in this area.