#143 Fighting the Climate Crisis with Data

Hi, I'm Maggie, VP of Curriculum at datacamp, and I know firsthand how overwhelming it can be to keep up with the rapid changes in data and AI. But staying ahead doesn't have to be hard. That's why our curriculum team has curated the best learning content to help you and your team stay ahead, covering everything from data and AI literacy for busy executives to cutting edge generative AI for developers, and everything in between. Whether you're upskilling yourself or your team, we make learning practical, engaging and flexible. It just works. Power your data and AI transformation journey on datacamp.com.

We live in a so exciting era regarding data and digital, where you have innovation in the data space. Meaning here I say innovation in the field of systems that can help you build data products or data solutions in such easy and powerful way. Quantum computing is on the verge of happening as well. We live in an era where we live these data revolutions and in Europe, organizations for the people who listen to us. I'm sure that you are witnessing that. Don't be afraid to go and act boldly trying to deploy the most massive energy that you can have in your space with your community.

Hello everyone. Welcome to Data Framed. I'm Adele, data evangelist and educator at datacamp, and if you are new here, Data Framed is a weekly podcast in which we explore how individuals and organizations can succeed with data and AI. Every year we become increasingly aware of the urgency of the climate crisis. With that, the need to usher in renewable energies and scale. Their adoption has never been more critical. However, as we look at the opportunity of leveraging data to make our energy systems more efficient and accelerate the adoption of renewable energy, there's probably no better player in the game than Engie. For those who are not aware, Engie is one of the biggest energy producers.

In the world and definitely one of.

The biggest in Europe. They operate in more than 48 countries worldwide and have committed to becoming carbon neutral by 2045. And if you're an energy company, that's a pretty big challenge. So we invited Jean Pierre Pelissier, the Chief Data Officer at Engie, to share his perspective on how data science and AI are integral to achieving Engie's renewable energy ambitions, and how he's approaching the overall data transformation at engi. Throughout the episode, we discuss in detail what the energy industry's value chain looks like, how data fits in the entirety of the value chain, common challenges data leaders can encounter in this space, how data accelerates the transition to renewable energy, how to best approach steering the organization's data culture And a lot more. As always, if you enjoyed this episode, do let us know in the comments and in ratings and on social. And now on today's episode.

Jean Pierre Pdice, it's great to have you on the show.

Hi Adele. Thank you for inviting me. It's a pleasure to be here.

Likewise. So you are the Chief Data officer@engie.ng has been making significant strides in harnessing data for value creation. It's been an interesting year for the energy market, to say the least. So maybe to set the stage for our conversation, how does data unlock value for ng?

And maybe I can start with an example of what happened last year. You are aware that last year in the world we had major disruptions on the energy markets because of the war in Ukraine and at ng, some examples show that data is not really on the way to become a key asset along with our industrial assets or talented teams. I will give you one example is that our trading activity, our trading teams were able to surf through those very hectic and volatile conditions of the market to manage risk and balance our portfolio between productions and retails to ensure that we serve our customer while being fully covered on the market. And they were instrumental in the good financial performance of the company last year. And it was not a given. We could see that some of our competitors had major difficulties, major challenges. Some of them even went bankrupt or had to request the backup of the government. So that's an example of how data is today leveraged at ng. Otherwise, NG is today a global company which is positioned across the whole value chain of low carbon energy. We work from production to retail, but we also store electricity and gas. We trade and we also offer energy solutions to customers so that they can optimize their consumption. I mean all of those activities. Today we have projects to deliver business value with data.

That's very exciting and I really appreciate that example and how you contextualized it with the disruption of the energy market and the war in Ukraine. You mentioned the value chain of energy. So maybe we're going to discuss a few use cases of data science and energy. Maybe it's good to also demystify what that value chain looks like. Maybe walk us through what the different sides of energy production looks like, what that value chain looks like, and where does NG sit in that value chain really?

So NG is a global player of this value chain and it all starts with production. So it can be production of gas or production of electricity. So we have assets of, let's say facilities which can be gas, thermal plants which uses gas to produce electricity, but they are the, let's say, the old way to produce electricity. And on the other hand, we now have assets of renewable energy that use solar or wind or biogas that can produce electricity. So that's one part. The other part is to be transport. After the, be it gas or electricity, we have assets to transport those components. It's a molecule in the face of gas or electrons for electricity. We can transport that on very long distances, from thousands of kilometers. And after, once they are transported on those very long distances, there are networks of secondary distributions that go to your house or industrial facilities where after that electricity or that gas is consumed. And so we have companies to produce, we have company to transport, we have company to distribute, and we also have after companies that sell the electricity or the gas. Each time they are separated activities. You have to understand that each time it's different entities, different business organizations. And that's, let's say, the primary value chain. And then as I said, we also have other activities which are complementing these value chains, like the storage. We have seen that in the winter in Europe and this year there were a lot of attention given to where are the levels of storage of gas. Do we have enough gas storage to go through the winter? And actually NG operates such facilities of gas storage. It's cavities that we have underground and where we inject gas when the consumption is not so high, and we consume the storage of those gas when it's needed, for instance, in winter. And finally, we have an activity of energy providing advice to industrial customers about their energy management. And we install some solutions to optimize the energy consumption and to reduce their energy footprint and carbon footprint also, by the way, and I would say finally, yes, there is the trading activity as mentioned it in the introduction, but that's also an important part of the company. We have about 3,000 people working in the training activity today. Global Eng is a 96,000 employee company and we are present in 31 countries today.

Thank you so much for this really detailed overview. And what's very interesting is how complex the value chain is. There's definitely a lot of different activities here. And moreover, leading data at a global organization definitely gives you a lot of unique challenges. I assume as a chief data officer, what do you think are the most critical success factors for organizations, especially especially in the energy sector, but especially in global organizations such as engie, when you want to harness data effectively for value.

Creation, Data is effective when it can be shared, when it can be easily accessed, and after it is computed to Create analytics and value. So I would say that what is really critical to unlock that is first to have a culture of collaboration in the company. And it's true that at NG we are coming from a culture of decentralization where the local organizations had a lot of autonomy. And so that is a situation where data is not easily shared between organization and entities. And so one of the critical part of our action is to ensure that now we make people understand that yes, of course they are perimeter to manage data, but they could benefit from data coming from other functions. If you are in the industrial production part, of course you need the finance data to operate, to understand how your P and L is going. But also it could be useful to have access to some HR database when you need some trained people to hire in certain fields. This is an example of to say that if we unlock the data, we create more value. Let's say it's a very basic assumption that I'm saying, but that's not so easy to achieve in a wide company. I would say that the very first thing is to ensure that this collaboration mindset, this data sharing mindset is there. And so after what we've done is that we've. The second part I would say is that we have infrastructures that are effective and easy to use. So that people when they are convinced and they are saying, okay, I'm ready to share my data, but if it's cumbersome, if it takes hours or it's to share it, I will not. So we have after to build easy to use infrastructures and I will. So that's the second part. And the third one I would say is that maybe the most important one in the end, once you have those fundamentals of infrastructure and mindset, is that you focus your actions only on the most value creation. Because otherwise you can waste your energy and efforts on things that are useful, but not providing the much value that you could, you can. And that's where having a strong governance is essential.

That's very great. And we're going to unpack a lot of these elements you mentioned here. A culture of collaboration and essentially this connects to a data culture within the organization. You also mentioned the infrastructure component, making sure the data is easily accessible, also prioritizing value creation use cases. So maybe let's focus a bit on the data collection side, on the infrastructure side. You know what's interesting and what's pretty fascinating about the energy sector is that the data sources that often are used to create value and analyze Data come from IoT devices, sensors pipelines that measure the effectiveness of pipes. Walk us through the data collection and quality challenges in this space. What steps has NG taken, for example, to ensure data accuracy, completeness? What does a data governance framework look like in this particular context?

One thing I would mention first is that our journey is clearly not finished in that space. We are in the middle of the journey. So we are working very hard to harness this data and have a complete view on the data collection and the data quality globally. But that said, yeah, if we go back to the value chain, you're right when you say that we have a lot of IoT devices that will send us data. We have IoT to send us data about how a windmill is working during all its day. We have IOT now. We have smart meters that are installed in many homes in Europe, in France or in Belgium, for instance, and they send you this information on how you consume electricity or gas at your home. And in the future, we will have a device that monitor how electric cars are consuming. So, yes, we have to deploy a very wide network of IoT to collect this data. And that was a tremendous effort. And in addition, you have to understand that we have a legacy industrial assets network which were not used for digital. So we had to deploy a lot of new tools or upgrades to be able to send the information in the end. So today, I think it's not complete. When we say completeness, we cannot say we're complete. But we have a fair view on how gas and electricity are now used at many in our production plants, but also at homes of our final consumers, but also industrial facilities. And with all this data, now comes a phase of, yes, are we governing this data well? Do we have enough quality? When there are quality defects, do we identify them quickly or automatically? These are many challenges that we face right now. This is work in progress. But once we have this kind of information, we will be able to unlock a lot of value, of additional value to, to the company and to final users. Maybe I will tell. We'll give you an example of that again. During this winter, we had to face some peaks of consumption because of days of cold. And we were able to forecast those periods of strong cold and to organize after campaigns, for instance, in France, to tell consumers, beware, next Monday it will be very cold. And if we want to have a collective effort to ensure that we go through this, we offer you a challenge is to say to limit your consumption below this threshold. And that was adjusted home by home. We knew that that home was consuming more or less. Let's say 100 of energy. And we were challenging those people to consume only 80 during that day. And we would reward the person with a financial benefit. It would give you a €5 or €10 discount on your next invoice if you're able to do that. But if we can do that collectively on many homes, it actually helped to balance the system, the global system and the level of country. And so you see, having the vision on how a home is consuming, we were able to target the companies very, very effectively thanks to our data tools and data systems.

That's a really great use case in a lot of ways. We all remember I live in Europe as well, I live in Belgium, the winter. Especially with the challenges that came from the energy disruption that we mentioned earlier in our conversation. There was a lot of conversation and chatter within the media, within government, about how we need to ration energy to keep the system globally balanced. And seeing data science providing value in that particular challenge is very useful. To put it in context, you know, you mentioned as well how there's a culture of decentralization and we need to be able to remove silos between data sets and provide access to data sets to the wider organization. This is oftentimes with organizations that have a history that is as rich as engies, there's often legacy systems, as you mentioned. What are best practices that you can share when it comes to approaching this challenge of bridging the gap and creating that integration between different data sets, different parts of the organization and modernizing these legacy systems?

Well, the question here spans a global data roadmap for a global company. This is a multi year work that you have to achieve. You cannot break the silos and work in an integrated way in six months. It's a multi year effort. But that said, the way we do it at NG is that we work first on the. It's not coming only from the data. To be clear, actually we are piggybacking on the global strategy of the company, which wants to be more integrated. And so it actually starts from that global NG strategy to where we went from about 25 business units, now only four global business units, showing that we are much more integrated. And we divested as well some activities which were not related to our main focus, which is low carbon energy, democratizing the access to low carbon energy. So having this global framework of the company is of course key because after our message is to say if we want to be an integrated industrial company, the digital and data space needs to reflect that, needs to be aligned. And so breaking the SILOS will be part of the strategy of becoming more integrated. Having this global strategy is key, but after it takes then as I said, we need to at corporate level provide the infrastructure, the data lakes and all the middleware effective enough so that people can actually break the silos while still retaining the ownership of the data and ensuring the cybersecurity of the infrastructure globally. And not least also to comply with privacy regulations for instance, and other laws which are national, even though we have a global platform, you need to comply with the national and local regulations. So here at corporate level we had to be very disciplined in understanding what kind of infrastructures could be really needed for the benefits of our entities or organizations. And then it's a lot of effort about networking, going to, working on the field with the people to co construct the needs regarding the infrastructure that will be used after by the entity and to promote the adoptions. So I would sum up by saying that yes, if we have a corporate strategy, it helps if you have after and then it's building an infrastructure that is co constructed with the entity and that requires a lot of going to the field, going to meet people, a lot of time and energy of course to enhance and adopt all those and make people adopting the infrastructure. So you, you, you cannot be sitting only here in the head office and expecting that everything will converge global infrastructure alone.

This is really great insight, Jean Pierre, and we'll definitely deep dive into what that engagement looks like with the wider population and how do you build a culture of engagement. But I think now this marks a great segue. You mentioned the use cases that you worked on when it comes to reducing global energy consumption within the system. But I think also what's really important to discuss is how data science unlocks value for the renewable ambitions of Engie. We discussed before our, our conversation today really how ambitious the renewable energy goals are for Engie. Maybe walk us through these ambitions and how does data science unlock value when it comes to the renewable energy transition?

Have to realize that Engie's ambition is to be carbon neutral by 2045, an energy producer and being carbon neutral, let's say in 20 years, a bit more than 20 years, it's a huge challenge actually. And, but it also shows the the power of the ambition and of the visions that we are all committed here in the community. And if we talk about renewable specifically renewable will focus the main part of our industrial investment in the coming years. And we plan to be at 80 gigawatts of renewables energy of production by 2030. So I mean at the end of this decade we want to reach 80 gigawatts of renewables. Today we are at 34 globally. So we want to add 50 gigawatts of energy, renewable energy production. You have to imagine that this is about 50 to 60 nuclear plants. It's the equivalent of the electricity production of 50 to 60 nuclear plants that we want to add by the end of this decade in renewable energy. And so clearly we have massive challenges in the building and the construction, in the project development. But once we've done that after have to ensure that the production is done in the most effective way. And this is where data science today is helping a lot. For instance to optimize the daily the renewable energy production. And I will give you example again. In the. In the solar field or in the solar field you, you, you are dependent on of course the weather forecast. And you need to. To adjust very, very accurately the orientations of the panels to ensure the maximum output possible. If you are not well positioned, you will not deploy that effectively. And now we have data science algorithms that can track where some that compares the actually that real, almost real time weather data and solar data that is coming in a specific region to forecast what should be the maximum output that you see. And if locally the plant is not reaching or is really far from that production, it will be detected by data science. And that may trigger some maintenance or at least an investigation of why the site is not producing as it should based on the observed weather conditions. And that's the same for wind. For wind, of course the production is dependent on the wind strength and on the wind directions. And again we can use data science to understand whether the windmills have produced the output that could be expected from the specific and local weather conditions. That's in the field of renewable how we are using today datasense after in the future we will use also datasends in fields of light maintenance of visual recognition because we can send our renewable assets are very often in remote places and they are quite difficult to access. And we can use now drones to. To inspect the farms and to collect the visual images. And visual recognition algorithms can help us detect some start some defects or some things that will need a maintenance. Having to send some technicians on site and it saves a lot of time for our teams, for our technical teams.

I think these are amazing use cases that you mentioned here Jean Pierre. And especially when we start thinking about the climate challenges that we're about to face in the next 20 years as well as food security challenges that we may face as well, when you mention these use cases, one thing that pops out to me is that a lot of these use cases require quite a lot of subject matter expertise within the energy space, renewable energy space, in many ways. You need to be able to marry data science skills with the subject matter expertise to be able to develop these use cases. How do you ensure that data science teams at NG develop the subject matter expertise? How do you marry that technical skill set with the knowledge of energy? And how do you build that out as a capability within Engie, what you're.

Saying is a critical part. Yes, data science. Our data scientists need to have a curious mindset to really understand actually the global value chain of where they are working to ensure that the algorithm they build are specific enough and really customized to the need of the operations. And we do that actually by really emerging the data scientists in the teams, in the operational teams. Of course we have a center of excellence which is central. But this center of excellence only provides, let's say, very first, initiate some projects or helps organization to think or to brainstorm and to start projects. But once we are in the real development phase, the data scientists, they are either internalized by the local organizations or they are sent for a long term mission on site. So we really have this mindset of emerging the data scientists in their as if they were other operational colleagues to the teams. So that's, to me, it's a very important part of that. And in the end, data scientists are scientists and I think that they also have quite naturally, I would say, a curious mindset because when you're a scientist, you want to understand and you ask the questions to understand. Most of the vast majority of data scientists I've seen or are actually very proactive in going to meet their colleagues and to ask them, okay, I will work on this field, but please explain me a little bit more. How does your device work? What do you want to achieve is really you would like to expect before they start to develop models or things like that.

So definitely learning by immersion is a great way to give that subject matter expertise. So we talked about how to give data scientists the subject matter expertise. Let's maybe flip it around here and talk about how to give the wider organization some data skills in the data culture. We mentioned mentioned it earlier in our discussion as well, how culture and opening up to collaboration and to data sharing is an incredibly important aspect about G's data ambitions. Maybe walk us through at the beginning, why is creating that data culture and.

Data mindset so important?

Right? And then let's discuss what are some of the initiatives that are done at NG to address that?

Yeah, it's a great question, Adel. And to me that's the. I think this is the most difficult challenge that we are going to face at nga and I think in every organization. In every organization, exactly. Is that if we really want to leave the digital era and the data era, let's say all the employees of the company should be at least a minimum level of awareness about how digital and data works and how it can create value. And then it means to have a little bit of understanding of what is statistics, what is the minimum mathematics that is behind if you really want to use effectively analytics tools even if they are low, code or no code. Today we know that we won't have enough people trained in computer science or in mathematics to really fill all the gaps of the harnessing data and creating value with it. So we to need need to upskill our workforce and it's going to be a massive need of training for all our employees. So that when you are an accountant and you use Excel, it's great. And you may be many accountants are very expert in Excel, but maybe they are not aware today that you have local tools that make the work they do in a bit cumbersome way, but expert way in a much easier way and unlocking a lot of other functions with advanced analytics. We need to make these people understand that those tools exist and train it to those local tools. And conversely, maybe there are some technician maintenance technicians on the field which have a low, let's say a low awareness of digital and data tools because they were not trained in those fields in the first place. But still in their private life they are using digital data tools so they know how to use apps. They know that somehow they understand a bit of the data sharing, what it means and so on in their private life. We need to upskill them and make them understand yes, in the professional life as well. We will provide you with app tools and data models that you will use to improve your work and the effectiveness in your professional job. Just give you this example. I'm quite sure that many people understand today how Netflix works and they understand that yes, if they look at this kind of movies or series, there are some algorithm that works in the background that after will prompt them with other series or films. So maybe they are not connecting the dots with machine learning and so on. But if you explain to them this example in their private life, you will will make them understand their professional life, how it works.

I couldn't agree More especially on connecting the way data is around us in our daily life and our personal life. Data literacy is something we all engage in and in a lot of ways, carrying that over to the professional life is a great message you mentioned in our conversation is you can't be a chief data officer that just sits at the head office where you need to expect things to happen. You need to talk to the organization. You need to talk to the people within the organization. Maybe walk us through. What advice would you give chief data officers when they're trying to engage with these rest of the organization and trying to build that momentum for a data culture and a data mindset?

So what we've done at Engie is that we have built a network. It's very important to build the network because of course, even if you cannot sit all day in the head office, you cannot be everywhere at the same time. So you need to have relays. And one of the very first things that we did is to build a community and to grow the community afterwards. And so that means, yes, working the data officers in the first place should start evangelizing, but trying to identify people that has this attractiveness to work for the data space and that will become data champions or local data officers. Sometimes I haven't done it for ng because I arrived only 6 months ago. But prior to my arrival, that network and that community was built over the last three years. And today we have a network of 40 data officers at NGE and almost 2,000 people which are, let's say, data friends. We call them data friends, meaning that they are people that have that in one way or another in their job, have an important data component. They are maybe data owners or they use some local tools like Power BI or DataIQ and they are part of our network. And that's. You see, after three years of work, it's about 2,000 people that we have. We are animating this network with monthly webinars and with some events like Data Science Challenge and some hackathons. So yes, there is a global community program around data that has been built at Engie and I would really encourage any data officers of a global organization to have that because this is how you can spread the culture and that you demultiply your energy in the company.

I couldn't agree more. This is really great insight. I definitely agree with the importance of creating a learning culture, the importance of creating an evangelist network within the organization. As you mentioned here, Data for Friends. This is something that we've seen also Data camp for business customers use Quite a lot as a tactic to engage the wider community. As we close out our conversation. Jean Pierre, what I'd love to learn from you as a data leader for many years now in multiple industries, driving value in organizations, in the space, whether in manufacturing, energy, et cetera, what would be your advice for leaders looking to gain an impact within the first six months on the job, within the first year? How would you structure that approach?

I would say that now we are in the phase where data is not just in the innovation phase. I mean, what I've seen is that in the years 2000, 2010s, digital and data were very often linked to innovation and it was like, let's be innovative and let's create a digital product. Now in the years 2000s we have gone out of the innovation space. Data is now a mainstream tool or a mainstream department that needs to be structured as professionally, as I would say an HR department or finance. Exactly. Data has to be a full fledged department in your company in the year now, from now on. Because if you don't do that, you will never unlock all the value of data at the level and the scale that you need to. And so to do that, what I, at least I share the way I've done it for ng, it's to go very quickly in building a roadmap to show how you will deliver business value with data. Because this is now what companies expect. It's no longer again innovation, it's business value. And your executive committees, they will want to see how data contributes to the growth or the EBIT growth. And you have to, to build a roadmap and a framework of actions around that. And once you have that, this north star of how do you create business value? Then you can have various flavors of actions. How do I govern the data so that the quality is strong enough to create business value? How is my infrastructure ecosystem going? Is it user friendly enough, is it flexible enough so that that anyone in the company can have access to it? And also in the culture part, how do I engage my, all the company, all my employees in the company to focus them on business value creation? With data, my feedback, I would say that it's again, it's never forget this north star of data is now here to create business value. It's not just an innovation or communication matter. So it has to be structured in a very professional way. And that's why you need the governance program, culture program and infrastructure program. And even in the end, a strategy or data strategy that matches the global strategic objective of your company.

That's really great insight, Jean Pierre. And maybe as we close out, one thing that you mentioned here that is key is choosing the right use cases that deliver business value. If you're a chief data officer that just joined a new organization, how do you prioritize those business use cases? What are low hanging fruits, what makes a low hanging fruit of a business use case and what makes a use case that is a more long term strategic objective? How do you prioritize and make the trade off between both?

As a data officer, you need to speak to other C level people and that across the organization, you cannot speak only to the IT family or IT and digital family because you need to have the attention of people that have the money to invest and to invest big in projects that will generate return on investment and that will help the P and L of their entity. So a lot of things that you do during the first month is to go to speak with C level people and their teams and their executive teams to understand what are the real strategic, what are the real pain points of the company? And even if it's not in the digital part, as a digital and data team, you have to work in framing projects or products that will help solve those pain points or achieve those business objectives. It's really spending a lot of time in talking with your business partners because I know that sometimes when you arrive as a data officer, you can be drawn in meetings with your fellow colleagues of IT or with vendors, or with the digital and data economy ecosystem. I would say. But be careful of not being disconnected from the real business objectives of your fellow colleagues in the business organizations.

I couldn't agree more. Now, as we close our episode, Jean Pierre, do you have any final words or call to action to share with our audience today?

First, I would like to thank you, Adele, for inviting me and giving me this opportunity to voice this. We live in a so exciting era regarding data and digital where you have innovation in the data space. Meaning here I say innovation in the field of systems that can help you build data products or data solutions in such easy and powerful way. Quantum computing is on the verge of happening as well. We live in an era where we live these data revolutions and in your organizations for the people who listen to us. I'm sure that you are witnessing that. Don't be afraid to go and act boldly trying to to deploy the most massive energy that you can have in your space with your community. It takes a lot of energy to deploy your change culture and to make people savvy with data. But it's so interesting and it creates so much value that it makes all this journey very rewarding.

That is very well said, Jean Pierre. Thank you so much for coming on Dataframe. Really appreciated your insight.

It.