Aggreko’s mature data science and machine learning platform allows it to optimize its supply chain and logistics and improve customer experiences. Still, over time, the growth of its core functionality had resulted in sub-optimal operating processes.

For example, the battery alarm process required several manual steps and manual intervention. This placed an additional burden on internal resources to ensure the end-to-end process was fully supported.

These issues are common in organically grown systems and are solved through iterative design and refactoring.

Adam Sroka, CEO of Hypercube, working in collaboration with Aggreko’s Data Science Team Lead, recognized the need for a comprehensive yet quick turnaround project to address some of these challenges. Automating the manual processes was necessary to alleviate the workload on internal resources, freeing them up for more valuable activities.

Aggreko’s mature data science and machine learning platform allows it to optimize its supply chain and logistics and improve customer experiences. Still, over time, the growth of its core functionality had resulted in sub-optimal operating processes.

For example, the battery alarm process required several manual steps and manual intervention.  This placed an additional burden on internal resources to ensure the end-to-end process was fully supported.

These issues are common in organically grown systems and are solved through iterative design and refactoring.  

Adam Sroka, CEO of Hypercube, working in collaboration with Aggreko’s Data Science Team Lead, recognized the need for a comprehensive yet quick turnaround project to address some of these challenges. Automating the manual processes was necessary to alleviate the workload on internal resources, freeing them up for more valuable activities.

With just 13 weeks to deliver the project, setting clear and actionable objectives was a priority.

First, we set out to optimize the orchestration of the Emissions Modelling notebooks. The documents simulate various emissions aspects to perform sensitivity analyses, scenario planning, and policy assessments to make informed decisions about emissions reductions, so they were a vital part of Aggreko’s operations to fix.

Efficiency was critical, and the data science team was too reliant on manual operations to run end-to-end processes. So, we looked to find areas where routine tasks could be automated and free up internal resources to focus on higher-value activities – improving productivity and minimizing the likelihood of human error.

Another aspect of the project was the parameterization of the emissions modeling process – allowing for greater flexibility and adaptability in the modeling process, making it easier to tailor the model to specific scenarios, locations, or conditions.

Finally, we set out to develop an Asset Monitoring prototype architecture – intended to monitor and oversee equipment within the organisation – the delivery of a new end-to-end alarm implementation and knowledge transfer to the Aggreko team, empowering them to manage and optimize their data science operations independently.

We broke down the project into four stages:

Onboarding and Planning. During the initial week, our consultants became acquainted with Aggreko’s team, gained access to relevant systems, including the alarm monitoring system and the Aggreko Global Emissions Method, and began reviewing relevant documentation. This period was critical for laying the foundation for the project’s success.

Emissions Modelling Automation. Over six weeks, we ran the iterative development to achieve the parametrization of the existing data pipeline.

Asset Monitoring Rearchitecture. Our primary focus was implementing new alarms while producing comprehensive documentation and templates for the Aggreko team.

Wrap-Up and Knowledge Transfer. Finally, we ran workshops with internal stakeholders to debrief on the engagement, share relevant learnings, and identify follow-on tasks, ensuring they managed and maintained the newly implemented processes autonomously.

Our strategic engagement with Aggreko exemplifies the power of orchestration and automation in enhancing data science efficiency. By addressing core challenges and prioritising knowledge transfer, the project achieved its objectives within the agreed timeline, setting the stage for Aggreko to continue to leverage its data science and machine learning platform to drive success for its clients.