Client Stories
Total Energies
Machine learning makes energy usage predictions more efficient
Wouldn’t life be better if you could make more accurate predictions based on data? That’s what TotalEnergies thought too, which is why they came to us to create an all-singing all-dancing machine learning tool to improve their reporting.
<>
TotalEnergies is a broad energy company committed to finding solutions to the challenges of climate change. They’re a real success story and their innovation has helped TotalEnegies withstand the test of time. Founded way back in 1924 to help France play a key role in the great oil and gas adventure, they’re now a booming broad energy company working on a global scale. We’re talking oil, biofuel, natural gas, green gases, renewables and electricity! The whole shebang!
<>
The challenge - the tech wasn’t as awesome as their aspirations
TotalEnergies has over 100,000 employees who are committed to making energy more affordable, more reliable, cleaner, and more accessible - which is incredibly awesome! But its internal data and BA (Business Analyst) teams had a problem. They wanted to use data to make better predictions and improvements across the 130 countries they’re active in.
Their existing data setup wasn’t very fast and frustratingly wasn’t scalable, which was making it difficult for them to reach their lofty goals. The good news was they had a pretty decent data infrastructure and adequate systems in place for BA reporting - but they wanted more. And we all know how that feels, right?
They wanted to be able to build predictive models around usage data, weather and demand forecasting - how cool would that be? But unfortunately, they were limited by their system as it took a really long time to run, it wasn’t scalable, and they couldn’t run parallel workloads. They wanted to be more efficient and more adventurous, which are two qualities we bloody love at Razor. We knew we could create something really awesome for them that would match the ambitions of their BA team.
<>
Action - try before you buy machine learning
We started this project the same way we always do, with one of our tried and tested discovery sessions. We explored their requirements, and priorities and spent time getting to grips with their existing processes and capabilities.
We joined this project early on, which is always really fun for us and means we hit the ground running. We did an assessment of Azure Databricks and Azure Machine Learning to look at their different options. We built a comprehensive (and pretty awesome) report on the pros and cons of each option so TotalEnergies could make an informed decision on the right tech for this project.
But we didn’t stop there! We set up simulations on both solutions with an example input - letting TotalEnergies try out both options - pretty cool huh?
Our assessment of both platforms left us with the conclusion that Databricks was the way to go, it would do everything TotalEnergies needed and more! Which would give them room to grow and develop, perhaps even inspiring them with new things to try!
<>
Results - machine learning is helping TotalEnergies stay ahead of the game
Want to know all the juicy tech details? We know you do!
We built a cloud-based solution that allowed them to build their own Python models to track data and where it’s coming from. We also embraced MLOps (Machine Learning Ops) so they could collaboratively build and run data models on the cloud for large compute power. Wowsers!
TotalEnergies are now set up on Databricks and are running their Machine Learning workflows in the cloud.
Do you want to explore how the Razor team and machine learning can help grow your business? Get in touch.
We built a cloud-based solution that allowed them to build their own Python models to track data and where it’s coming from.
