ArticleOctober 2, 2023 · 3 min read time
Virtanen.ai is an artificial intelligence application that indicates where electricity is used in one's household. It provides insights into how much people could save by scheduling power usage during cheaper times and what the amount of carbon dioxide emissions produced by personal electricity consumption looks like.
"We launched the service at the end of 2022, and so far, it has been utilised in analysing the electricity consumption of about 4,000 households. Throughout 2023, we have also introduced new features, such as the carbon dioxide emission calculation, which is not available elsewhere," says Esa Puttonen, the Data Scientist responsible for the service's development at Nitor.
The free, open-to-all service links household consumption data from Fingrid's open data hub with information on electricity market prices, weather observations, and electricity emission factors, delivering a precise view of individual household consumption.
"In the fall of 2022, when electricity prices rose, our analysis indicates that households reduced their usage of large electrical devices by about 20% compared to the previous year. Large electrical devices include sauna stoves, water heaters, and electric cars," adds Puttonen.
Virtanen.ai, designed to be as user-friendly as possible, is available in Finnish and English. Users must retrieve their home's consumption data from Fingrid's online service to leverage the service. Virtanen.ai guides users through this process, computing an overview in just a few seconds. The resulting report can be easily viewed on a monthly or yearly basis.
The pro bono project, developed during working hours, seeks to reveal household power drainers
Virtanen.ai was developed during the so-called "Nitor Core" time, within which Nitor employees can dedicate their working hours to internal development projects and their personal intriguing ventures. The motivation behind developing this service was to assist households in understanding their electricity consumption, as many lacked the tools to monitor their devices.
The energy crisis and resulting spike in electricity prices have piqued many consumers' interest in spot prices and monitoring electricity consumption. The reports provided by electricity companies display consumption data at best daily, making it challenging to gauge the power consumption of individual household devices.
Virtanen.ai breaks down total consumption by device group and calculates the costs of using different-sized devices based on electricity market prices. In turn, the emission data for different device groups are based on Fingrid's open market data.
"We don't know the exact devices present in households, but we can estimate the size categories of devices from the spikes in overall electricity consumption. We do this quite accurately, subtracting base consumption and heating before making the estimate. For instance, from the analysis results, we can observe that small electrical devices hardly matter if 80% of the total consumption is attributed to heating and sauna usage," explains Puttonen.
The development of Virtanen involved data analysts working with artificial intelligence, user interface designers, and software developers. The first version of the service saw the light of day in just a few months, but work on new features continues. Now, users can share their analysis with family members, for example, or save it as a browser bookmark, which means consumption data only needs to be uploaded to the service once.
Virtanen.ai is hosted on the AWS cloud without dedicated server resources, allowing it to scale automatically based on usage. The electricity consumption analysis was developed using the Python programming language leveraging data science libraries. The web service, in turn, was developed as a responsive web application using the NextJS framework.
Find out more about our various Core projects.