Technically Speaking

Climate change is a pressing issue and we need to learn more about how our actions affect the environment. There's a lot at risk but policy and regulation can help us change our impact on the planet. But to create effective climate governance, we need tools - to help manage compliance, and data - to validate the impact of our net-zero efforts. This leaves us with the question of what role technology, especially open source, can play in slowing climate change.

00:29 — INTRO ANIMATION

To achieve net-zero emissions, companies will need to minimize their environmental impact. Climate risk analysis and environmental, social, and corporate governance can help but there are still major obstacles to overcome. Integrating climate considerations into investment processes may be one way we can enable businesses to achieve their sustainability goals. But why are we even looking to the financial sector? To start with that why, let's talk to Cara Delia about OS-Climate, an FSI initiative at the Linux Foundation that's looking to address these challenges. Hey Cara, how are you doing?

Hi Chris. I'm great. Thanks.

So let's just jump right in. Why would financial institutions even want to address this problem in an open way?

Financial institutions are obviously heavily regulated and to get ahead of these regulations, they want to help formulate and influence the ESG reporting structure by working with regulators and with their peers in financial institutions.

When you think about this Regulatory Compliance space, this isn't really where open source is known. So, how do you see getting these, these different companies, even competitors, coming together to work on a common problem? And what role is open source playing here?

Well, this is the perfect opportunity for open source, because through regulations of FASB, through SASB, through ISSB. There are, and UN regulated, or UN sources, the financial institutions are having to do this. They're having to create plans for transition risk that help support the Net Zero Alliance, because the real issue within ESG reporting is being able to gather, as you said, the disparate data sources and really this is about any industry that you're in, and OS-Climate helps to address this.

So it sounds like there's some cool technology under the hood here. I know having a common platform to bring data together to look at federating both proprietary and open data sounds like an interesting topic. Where should I go to get more information on that?

Well, the beauty of open source is that the contributions can come in a non-technical and technical fashion, and I am not the technical person. So a great person to talk to would be Vincent Caldeira. He has been the key creator of the Data Commons platform, and he is a wealth of resources.

Thank you so much Cara.

Absolutely. Thanks Chris.

Hey, Vincent. I was just talking to Cara about OS-Climate and the work that we're doing to bring data together to solve these really significant challenges across the industry for understanding the carbon footprint of a business and she directed me to you to to help dig into some of the details of what it means to bring data together on a common platform and do data science in this OS-Climate context. You want to give us a little bit of the first view of what that means?

Absolutely, Chris. So from a data challenge, we are actually trying to solve mainly two issues. The first one is about data availability and how to make climate data easily accessible to data scientists and data engineers that are building climate risk model. The other aspect is data reliability where traditionally it's very hard for data engineers to assess the data quality and the provenance of data from the sources that is made available to them. And for this we use a data-as-code approach. We treat any source of data, any data transformation as a piece of code that is version and managed.

I love that view of data-as-code. I've thought about this analog a lot myself– of understanding source code to binary artifact as a pattern that we can apply in data as the source and machine learning models as a binary artifact. I know OpenDataHub is a key part of the underlying reference architecture. How do all these pieces come together in OpenDataHub for OS-Climate?

We use OpenDataHub as our baseline and blueprint for building a self-service data science platform. OpenDataHub by itself doesn't address some specific challenges that you would typically find in an enterprise context. One example of this type of challenge is around data governance and data security. We have basically started to identify open source projects that provide additional capability, such as, you know, metadata management, ability to tag sensitive data and the ability to manage code based access control around data in a consolidated and standardized manner. So we identify such open source projects that we feel have a wide adoption across the community and then if we deem them suitable, our job is to of course teach them together back into these OpenDataHub blueprints.

I like the comparison of what you're doing here to needs in the enterprise. So while this is really important for climate There's also the federated view of data, how you can write access to data without bringing it all together in a common data lake, as well as governance. These are these are issues that go beyond just OS-Climate. And then if you think about the underlying tools and bringing data together, Kubernetes becomes a key part of the story. I think it fits really well. How do you see Kubernetes being a key ingredient in this OS-Climate context?

A lot of use cases that we actually dealing with, may require real-time data sources,

volitional database and sometimes data-at-rest to be aggregated in real time. So what we found is the ability to run a distributed query engine on Kubernetes, to virtualize the query of the data across multiple sources is actually an extremely useful component for data federation. So that is very much a characteristic of Kubernetes we leverage for the platform.

This has been an awesome view of what's happening in OS-Climate, critical for understanding the carbon footprint of businesses, but also addressing these broader challenges and availability of data, bringing federated data sources together, the governance associated with that, even the kind of matchmaking of workload to underline resource. They had a GPU or accelerated environment. This has been really useful, Vincent. Thank you so much for the time.

Thank you, Chris.

It's not enough to just label companies or Investments as green or not green. It's a nuanced process of evaluating disparate data sources and measuring the potential impact of climate related issues on an asset’s performance. Equipped with these insights we can make wiser environmental decisions. Open source data-driven platforms such as OpenDataHub can give us a roadmap that, when paired with upstream communities and strong climate governance, will spur further progress towards a greener future.

08:03 — OUTRO ANIMATION