Committed to energy transitions, the French electricity grid manager wants to expand the use of data in its regional divisions to drive internal innovation. The program aims to open up the 10,000 Enedis data sources to a much wider population than data workers thanks to the native combination of Power BI and Dataiku.
WITH 1.4 million km of transmission lines, 800 places AND 35.6 million meters, Enedis manages the largest electricity distribution network in Europe. A network that today is concerned about new environmental challenges, for example the arrival of new energy sources or the evolution of use, especially the increase in the number of charging stations for electric vehicles.
Data will play a key role in this transition. Enedis already has about a hundred data scientists working specifically on the network load curve and on producing open data for communities. The company must now go further: “ Two years ago, we wanted to share this information with the entire company and our employees, explains Benoît Locu, head of Enedis Innovation Labs. This data is necessary for our transition. We want to open it to as many collaborators as possible, and not just to a lucky few. »
Opening data to regional employees and managers
Therefore, an extensive program is being launched in 2022. Appointed for regional administration, that’s what it was called DataDR. It is based on four main pillars: a unified and simplified data access process, comprehensive mapping of data sources, compliance with compliance and finally providing data manipulation tools that can actually be used by everyone.
There cartography the first to discover no less than 10,000 different business establishments. Faced with such a variety of data sources, the team had to prioritize those that would be open to users. ” We decided to adopt a use case based approach. This is not as intuitive an approach as it seems because in the data world we often think about making everything available before we see what we are going to do with it. However, if we wait until we can provide access to our 10,000 objects, it will take a long time and potentially be very expensive. Rather, the idea is to start from concrete cases and directly target the right data for the right people, create value and create a snowball effect. »
Another key point of the program was the selection of tools that would enable users to use this data on a daily basis. ” The DNA of the project is its “for everyone” side. We cannot impose language Python and make overcomplicated items available to everyone. Anyone should be able to manipulate this data under good conditions after a few hours of training. »
The project team decided to rely on two complementary tools: Microsoft Power BI and platform Dataik. ” These solutions rely on other software blocks that retrieve data from our information systems, our Hadoop data lake and our Teradata data warehouses. Specifically, we are implementing a data virtualization solution Denodo. However, it is this part of the user interface that remains the most important in the architecture, as it is what will make data access work for everyone. »
Enedis chose Power BI and Dataika to provide access to internal data for all its employees.
Power BI is the preferred tool for creating data visualizations.
While Dataiku is considered a low-code/no-code data manipulation tool.
Choosing to implement the system according to the use case quickly bore fruit. Data DR now has around 300 daily users, with a new case processed every three days. Among them are predicting the load on the Enedis reception centers, which process eight million customer contacts per year, or managing actions to connect new subscribers. “ Since 2022, we have implemented about a hundred initiatives and made several hundred business facilities available. Even without reaching the total figure of 10,000, we are already producing value. »
25 Enedis laboratories are involved
To set up the program and then discover these use cases with regional management, Benoît Locu relied on the existing structures at Enedis, especially the experience and tools implemented by the digital factory.
On the other hand, he could call online from 25 Enedis Labs, innovation labs spread across France. ” We asked them to record use cases and go back and forth between us, who provide the services and data, and the users in the regions that need the data. So we have data experts involved in the process and we have quality contact with the regions to detect these cases. »
Benoit Loc
Head of Enedis Innovation Labs
“We were looking for low-code/no-code solutions so that all employees could easily manipulate data and become data citizens. »
After a year and a half of work, Benoît Locu draws up an initial assessment: “ The tools were praised. We wanted to find those that were adapted to our employees, easy to use, reliable, and that goal was achieved. We have very good feedback from our users and it is a real success. »
The manager is pleased with the dynamics of the community, with hundreds of people using the solutions every day.
On the other hand, as far as open data is concerned, there is still progress to be made, and the goal of 10,000 data sources is still a long way off: “ We still lack important data to collect. This prevents us in certain use cases, we have to search for them. »
Furthermore, integration difficulties and the cost of connectors call into question the relevance of connecting end-of-life applications that will be replaced.
Finally, Benoît Locu wants to improve the speed with which he makes new data available. ” We have set a deadline of two weeks for the delivery of the data requested by the user. In certain use cases we manage to maintain this KPI, but in others it can reach from three weeks to a month, sometimes up to three months. We still have to make progress, because even a short deadline contributes to user satisfaction. »
Benoît Locu is also working on improving the positioning of Data DR in relation to national information systems. Indeed, we now have to decide whether the analytics project should be performed by one of the existing business intelligence systems or whether it will include Data DR. Each project must be subject to arbitration. Realigning each person’s scope should allow the whole to be collectively more effective.
Project figures
100 cases treated from use
300 users Daily
150 business facilities made available, i.e. 500 attributes
Enedis company
Activité : Gestion de réseau électriqueEffectif : 39 200 collaborateursCA (2022): 15,2 Md€
READ ALSO:
READ ALSO:
