Our secret master plan (don't tell anybody)
By Matt Gray, Co-Founder and CEO
As many know, our initial product was computer vision algorithms to estimate productivity and emissions from large fossil fuel facilities in regions where data is unavailable or infrequent. However, not everyone knows our goal is to create a global standard for energy planning data. This is because our overarching purpose is to help ensure affordable and dependable clean energy for all, especially in low-income countries.
Critical to making this happen is accessible, auditable and reproducible data. This is why we are developing Model Builder, an energy transition data platform providing net-zero scenarios for over 160 countries. Even so, some may question whether this does any good. Do we need another data initiative?
The answers to these questions are not much and no. However, that misses the point unless you understand our secret master plan alluded to above. Ambitious and urgent policy changes are required before clean energy is affordable and dependable for everyone. Those changes are most likely to be implemented in an environment where a wide range of stakeholders are putting evidence-based proposals forward.
To date, products to support energy transition planning are often based on closed data and models. This situation makes it nearly impossible for all stakeholders to engage in the conversation productively. The result is linear and incremental policy development, particularly for low-income countries, where governments often do not have the resources to afford expensive consultants.
To solve this problem, we plan to build an open-source energy transition data platform without usability constraints and partner with organisations and initiatives that help scale our user growth. More specifically, our strategy is to take the most promising open-source frameworks and:
Improve the quality, coverage, and provenance of their input data
Reduce usability barriers by modernising both programmatic access and no-code user interfaces
Create a network effect by partnering with users on both sides of the ledger
We intend to utilise OSeMOSYS and PyPSA, which are already among the most widely used open-source energy modelling frameworks. For example, these frameworks have been copied by well-known consultancies and initiatives (e.g., McKinsey, UN PRI), used as a core planning tool for several governments (e.g., Cyprus, Costa Rica, Bolivia), and incorporated into major technical assistance initiatives (e.g., UN DESA and CCG Programme).
Now let’s dive into the above strategies one by one.
Improving input files
Energy system modelling has historically been a victim of “garbage in and garbage out”, the concept that flawed input data produces flawed outputs. There are many reasons for this, such as the sheer number of assumptions, the confidential nature of energy infrastructure information, and the purpose of least-cost modelling. In short, for intentional and unintentional reasons, the outputs of energy system models are often detached from what happens on the ground.
Data science and market expertise can help overcome these shortcomings. For example, our Coal Asset Transition (CAT) Tool provides facility-level estimates of productivity, emissions and power purchase agreements (PPAs) from our detection algorithms and in-country expertise. The model results from CAT are then fed back to Model Builder. The PPA data for coal power plants, for instance, is hard-coded into the current policies or ‘business-as-usual’ scenario in Model Builder. The provenance of all data is traceable back to their original sources, and users can swap out data and assumptions suiting their use case. The outcome is a better understanding of how an asset's economic and financial performance affects system performance and vice versa, or more plainly: a more helpful product for energy transition practitioners.
Modernising the data stack
A data stack is like a kitchen for data. Think of how you bake a cake. The ingredients become a cake after going around the kitchen. Most cake ingredients are inedible by themselves, but after spending some time in the kitchen with a cook, they are transformed from inedible ingredients to a cake. And that is the key: unstructured data are unusable, like cake ingredients, are inedible. But after journeying through a data stack, this data gets turned into insights, easily digestible by various stakeholders.
The technical expertise required to use the data stacks underpinning energy system models tends to be with consultants and academics, who are not incentivised to be proficient in software development - a requisite to making a modern data stack that reduces usability barriers. To continue the cake analogy, expecting a wide range of stakeholders to use energy system models based on current data stacks is comparable to expecting my cake-loving four-year-old to bake a cake from scratch.
Our platform will allow non-technical users, such as policymakers and analysts, to quickly design and run models that project how power production and generating capacity will grow in the future, using data already on the platform or by bringing their own. These models can explore different scenarios, including climate policies, technology costs and pathways, and infrastructure investment. Models can be shared in various formats with colleagues and partners or kept private and confidential.
There will be some stakeholders, of course, who will want to tweak the recipe before they order it. Others will just want the recipe card they can use in their kitchen at home. Our data platform allows cake lovers of all cooking abilities to get the cake of their choosing.
Scaling users
With Model Builder as our data platform, we hope to create a network effect through a scalable partnership model. A network effect is where the value of a product or service increases as more users engage with it. Take climate finance as an example. Currently, high-income and low-income countries have no shared understanding of how much and what type of finance is required.
Beyond negotiation tactics and political realities, part of the reason why climate finance is so fractious is there is no global data standard - no shared understanding of how much ‘net zero’ policies cost beyond high-level reports based on closed data, which cannot be audited or replicated by all stakeholders in low-income countries.
One of the reasons why a global standard for energy transition data has not emerged is due to hoarding behaviour to preserve business models and avoid credibility challenges. We’re building open source, anyone can use the model without restrictions, meaning the more users we can attract from host and donor countries, the more accepted and powerful the data will become.
We plan to attract users from low-income and high-income countries to generate a network effect and support the development of a global standard for energy transition planning data. For example, through our partnerships with V20, ESCAP and CCG, we provide data to support decisions to increase climate and concessional finance. If host and donor countries increasingly rely on the same data, then it seems logical to assume data will be less of a blocker in preventing the flow of climate finance to low-income countries.
Owning the ecosystem
In short, our master plan is to:
Build an open-source energy transition data platform without usability barriers
Create a user network effect through a scalable partnership model
Become the industry standard for energy transition planning data globally
In an increasingly open-source world, owning the ecosystem is the only way to win. By owning the platform where innovation happens, we can cement ourselves as thought-leaders and direction-setters, allowing us to shape the narrative on ideas larger than ourselves. In doing so, we can play our part in making clean energy affordable and dependable for all.
Don’t tell anybody.
