Attiki Gas Supply Company (EPAA) is the single major gas supplier in the Attica basin, serving 330.000 household consumers, 200 large commercial and industrial clients, 6.500 businesses and 1.200 schools and public buildings.

Part of its obligations, as a client of the National Natural Gas System (NNGS) run by DESFA, is to report on a daily basis its natural gas capacity booking. This parameter is based on the daily natural gas consumption estimates, which the company acquires applying machine learning techniques and utilizing weather forecast data. The NNGS regulation requires that these estimates are as close as possible to the actual natural gas consumption by its clients, liable to penalty fees for positive or negative deviation. In short, the accuracy of this estimate is directly related to the company’s cost savings strategy.

Ex Machina worked closely with EPAA to address this challenge. The first step was to supply the necessary historical weather data for training the machine learning algorithm. A set of weather data time series from various locations were used, in order to locate the most appropriate points of interest. Once these were identified, the next step involved the automatic delivery of the data duly processed and in the desired format. After validating quality of service and accuracy of results, the Ex Machina Weather service started feeding the production machine learning system of EPAA and saving them money!

Since 2016 we’ve had a very close collaboration with EPAA. The few times that was required from us to adjust our service it was done almost instantly and transparently for the client. The added value of our service lies in the flexibility of delivery (web service, api, email, etc), the swiftness of execution, the ability to request for custom data processing prior to delivery, the quality of data and the agility to experiment with diverse models and approaches.

But why not let our customers say it for themselves?