Orchestrate all the Things - Graphs to trees: data, metrics, analytics and regulation for sustainability. Featuring Neural Alpha CEO / Founder James Phare

Episode Date: June 23, 2022

What does sustainability actually mean for organizations? Can it be measured, and if yes, how? Obvious questions with less than obvious answers, even for sustainability and ESG professionals like James Phare. Phare shares his experience and assessment of sustainability and its relationship with the ESG space, its current state and trajectory, and how data and analytics can help.

Welcome to the Orchestrate All the Things podcast. I'm George Amadiotis and we'll be connecting the dots together. What does sustainability actually mean for organizations? Can it be measured and if yes, how? Obvious questions with less than obvious answers, even for sustainability and ESG professionals. We take a look at sustainability and its relationship with the ESG space, its current state and trajectory, and how data and analytics can help.

I hope you will enjoy the podcast. If you like my work, you can follow Link Data Orchestration on Twitter, LinkedIn, and Facebook. I'm James Fair. I'm the CEO and founder of Neuron Alpha. We're a sustainable fintech company based in London. I also have another hat, which is how I know George very well, which is as the original founder and now co-founder of Connected Data World back in 2016. And we've had many fantastic years working together, they're pulling together various

connected data tech conferences and other activities over the years. In terms of my background, I guess, prior to Neuron Alpha, I spent most of my career working in financial services and kind of related areas. So I started my career at a place called Reuters Data, which became Thomson Reuters and then Refinitiv, essentially working in kind of data quality management and data reconciliation and related things within a kind of variety of financial instruments so looking at things like commodities pricing data and instrument reference data and various other things and that kind of really piqued my interest in financial data and got me really excited about the area. So then, yeah, then I moved to London and joined a large hedge fund company called Mangroup, who are one of the largest hedge funds in the world.

We had kind of various roles there, culminating in being group head of data, where I was leading various kind of data strategy initiatives in different areas of the business, getting things like data quality management, metadata management, data governance, stuff like that off the ground. Then, yes, left man group in around 2012, I think, and then started the journey towards being an independent consultant, working in the financial services industry predominantly. Spent a bit of time in investment banking, places like Commer Commerce Bank and HSBC and also various asset management businesses, places like M&G and Schroeder's Fidelity. areas. So a lot of time advising businesses on how to more effectively manage data as an asset, how to bring in, you know, better data governance policies, how to proactively manage quality. And I guess also, you know, coming out of all of that, better deliver analytics. So a lot of the time was spent, you know, helping to implement things like data warehouses and business intelligence solutions,

particularly in response to a lot of the regulations that were coming out of the kind of post 2008 financial crash. So looking a lot at things like EMEA and Dodd-Frank regulations around things like derivatives exposure reporting. Also spent quite a lot of time in working in financial crime. And I guess that was particularly an area that gave me a lot of exposure to kind of graphs and connected data technologies, using those kinds of things to you know perform more advanced and deeper kind of know your customer and and anti-money laundering capabilities and then yeah spent a bit of time outside of the financial industry working in government again kind of focused around similar things around data strategy and then eventually

came came back and and got to work on one of my first kind of ESG sustainability focused projects which was around 2016 when I got the opportunity to work with a large renewable energy investment manager, helping them get their kind of data strategy for managing a lot of their kind of real asset investments in wind and in how and and what what the opportunities were around applying connected data to the whole ESG space and I guess that's really been my focus for the last six years with with Neuron Alpha since then so we we're a software and data products business focusing primarily on applying connected data technologies to ESG issues. A team of around 20 of us, mainly data engineers, full stack developers, also some product management business analysis staff as well, mostly based in London,

or we have some staff overseas in different locations. Okay, great. Great, thanks for the introduction, James. And one of the things I was going to inquire about, and I think you briefly touched on that, even though in some ways you may think it's kind of obvious is what actually attracted you to to work in the ESG space. Well if you'd like to just expand a little bit on that before we actually go into some of the specifics on ESG

before we get to cover the technology because I think for the people who may be listening and also for the write-up, it's important to understand what ESG actually is and what are the blind spots there. Cool. Sure. Yeah, so I guess what motivated me, I mean, I originally started to study a lot about climate change and environmentalism at university. So I did a degree in economics with a big component being environmental economics and, you know, looking at things like negative externalities and Pigouvian taxation and all of these things.

And I found it really exciting, you know, how policy makers could use tools to try and, you know, make a more more sustainable place and then I guess I kind of entered the world of work and thought well that that's been really interesting but I'm not sure I'll ever get the opportunity to actually work in that that space and then was yeah very fortunate really that ESG has has become this huge mega trend within within finance and, you know, a lot of demand now for new tools and new data sets in that space. So that's that's really kind of driven that focus over the last the last six years, really, years really in my career. And then I guess yeah you touched on the the elephantine sized question of what ESG is and whether I can explain it which question than many would have you believe.

I think it's pretty hard actually because it's one of those terms that's really an umbrella term and even in its definition by the book, let's say, it really jumbles a few things together. So ESG stands for, let's state the obvious, it stands for

environmental, social, and governance. So in a way, you could say like, okay, it's all the stuff that doesn't fit anywhere else. So it's the non-hard financial metrics in a way that apply to organizations so and that's that's kind of the first thing that I find personally you know not not having a background in that and not having any real exposure to that space, I find that kind of impressive. So why lump all of those things together? Wouldn't it make more sense to actually have separate criteria for environmental and then social and then governance? Yeah, I mean, you know, sustainability, I think, is a kind of umbrella term, if you like,

which kind of bonds all of these things together. And I guess sustainability as a concept has existed for a long time. We're recently celebrating 50 years since the famous Stockholm Environment Summit in 72, where that was kind of seen as a real milestone in advancing sustainability as a defined concept. In terms of where ESG fits into this, I guess ESG has been born more out of the financial industry, particularly out of the kind of asset management side of the industry and it was as you as you exactly point out George it was kind of developed as a term to try and capture and shape and a framework to managing kind of all of these fuzzy sustainability focused metrics and particularly non-financial metrics which impact on the

the performance and the reputation of of of a company so yeah I mean in a in a nutshell ESG is very simple I mean it just stands for environmental, social and governance. And it's essentially within the world of finance. It's the kind of metrics and the data and also, I guess, the way that products are developed to address different considerations within those those three pillars so for example within the the kind of environmental pillar kind of common measures that you see there probably carbon emissions scope one scope two scope three emissions are the most kind of dominant focus within that space although there are other kind of considerations things like biodiversity nature loss and so on and so forth social tends to focus more around things like development sustainable development goals gender equality labor rights those kinds of issues and then I

guess governance is more tends to be more focused around the kind of traditional corporate governance which you know long before ESG existed was was a big focus particularly for active investors within the financial community who are typically engaging with their portfolio companies, voting at AGMs, advising them on corporate governance best practice. And that could range from anything to how companies are legally structured to the composition of their boards down to you know things like how they structure the share classes for bringing in external investors a whole whole different range of

of criteria so I guess in in some ways ESG is a very simple concept but then I guess when you start to look beyond the theory into the practice for how ESG is actually used by the financial industry and I guess the results that we see within the products the financial products that we buy it often becomes a lot more kind of nuanced yeah and I think that's that's quite an interesting area lot more kind of nuanced. Yeah, and I think that's quite an interesting area, particularly a kind of hot area of debate at the moment, I think. Yeah, and I think, well, you did mention the magic word, so metrics. And I think what many people don't actually realize is that, well, it does make sense to have some sort of metrics,

like some sort of ground truth, let's say, by which to navigate and to sort of be able to evaluate and compare. But what I think that many people don't realize is that it doesn't look like there's an agreed upon set of metrics at the moment. And that leaves room for, well, speculation and what has come to be known as greenwashing. And I think there was like a recent incident with Deutsche Bank, if I'm not mistaken. So basically, what happened was that the German regulators raided the Deutsche Bank offices, accusing them of greenwashing. So basically, labeling some financial products as being ESG compliant, when in fact they were not. exposure to this metric, to this, let's say, lack of commonly agreed metrics. Some years ago, I was very curious about the way that CO2 is measured and, you know, how the institutions that give CO2 certifications function and so on. So doing a little bit of research, what I came to realize was that

that doesn't seem to be like, again, a commonly agreed set of metrics on how to measure CO2 footprint. And the institutions that gave CO2 certifications are quite opaque about the way they go about it. So, you know, when we talk about things like evaluations and data and metrics and analytics and all that, that seems sort of, you know, like a non-starter. So how can we make progress if we can't measure, if we can't even measure those things? Yeah, I mean, I think you've hit on a couple of really interesting points there around um kind of standards standards convergence and things that are happening there and um yeah

and also kind of voluntary markets and the and the risk of greenwashing um yeah i mean in terms of on on on the greenwashing side i think it's it's a really interesting time at the moment, as you point out. And, you know, greenwashing, you've had this huge growth in terms of profitability and assets under management within the financial industry. You know, it's something like nine out of 10 dollar inflows into US funds at the moment are into ESG funds so you've seen this huge huge growth of ESG products but now the inevitable kind of other side of the the sword is is coming to bear in terms of the risks of greenwashing because I guess a lot of the financial industry have been kind of really rushing to keep up there's been this huge war of talent and and we know you know it takes it takes

a long time to develop really credible detailed infrastructure to actually manage the kind of ESG aspects of your your portfolios and your your your loan books and this kind of side of things. So inevitably, yeah, now we're starting to see the kind of early greenwashing cases and, you know, regulatory sanctions, but also litigation being there by the German regulators, BaFin, emerged after an internal whistleblower blew the whistle, essentially saying that DWS were not actively incorporating ESG into their products, partly down, I believe, to use of kind of legacy technology, making it difficult to actually incorporate this data into their practices. And that was misaligned with their marketing, where they were marketing these products as being ESG products products when essentially they weren't kind of fully complying with those standards. So yeah, so I think, I mean, you touched on, I guess,

the example of carbon markets, but it's still an evolving space. There's many, many different standards out there. Obviously, I think we're going through a period of convergence where we will end up with less standards, but it's still fundamentally a voluntary market and you'll obviously see greenwashing and other risks. It is changing with the regulators becoming more involved in this. Maybe I could touch on some of the things that are happening there, which are quite interesting in a minute.

But I think the other thing that you touched on that's quite interesting is this kind of concept of standards convergence. So at the moment, it's quite a fragmented landscape and there are many different standards bodies out there doing different things around esg but there is a big uh kind of groundswell at the moment going on where different groups are coming together and starting to form coalitions to try and essentially pursue a kind of universal set of ESG standards particularly focused around producing universal ESG scores

that are comparable across different industrial sectors so you can for instance, compare the ESG score of Tesla to the oil majors consistently, which has been another area of controversy recently, if you followed some of Elon Musk's tweets, where he was complaining that S&P had rated a lot of the oil majors people like exxon mobil with overall higher esg scores than tesla uh mainly down to uh stronger governance uh sub scores which um you know he um he understandably was frustrated with um so yeah so there's a few there's a few interesting things going on in that space you've got got initiatives like GRI and Carbon Disclosure Project, CDP, and also accounting standards bodies, people like SASB, the Sustainable Accounting Standards Board. board, also other standards bodies, people like the Chartered Financial Analyst CFA Institute,

which is a kind of professional accreditation body, also working with some of these other bodies to try and produce common standards. But there are several groups out there, and I guess it's kind of similar in some ways to the parallel with the kind of VHS Betamax battle in the 1980s. So it's a bit unclear who will necessarily win out with those battles. But yeah, certainly we're in a period of convergence at the moment. Okay, so one way to think about it actually, I would say is, I would draw a parallel to what has been happening in data privacy. Because again, that's an area in which, you know, there was lots of talk and lots of different initiatives. But what really made the difference was when the EU just went

ahead and implemented the GDPR regulation. And that sort of pushed the needle across the globe, really, because everybody, so obviously, you know, if you do business in the EU, regardless of whether you're based in the EU, you have to comply with that regulation. So that really set the tone and advanced things, not just in the EU, but worldwide. Do you think that something like that, not necessarily coming from the EU, but do you think that somebody taking the initiative and sort of going ahead with regulation and a set of metrics would help in clearing the landscape? Yeah, I think the regulators are taking a lot of action and they're signaling

that more action will be coming. I don't necessarily think it will be a magic bullet. You know, when GDPR came in, I think a lot of people were understandably excited and also curious as to how it would change certain business models. I mean, I still get spammed by recruiters multiple times a day. So in many ways, it's improved things, but it hasn't fundamentally changed things. And I think there'll be a lot of parallels there with the experience. If I may say so, I mean, I'm sure you would still be spammed even before GDPR was in place. The difference now is that, well, you can actually do something about it, I guess, if you really want to.

That's true. Yeah, that's true. If you want to spend the time reporting all the different emails that you get. Yeah, absolutely. Yeah, so I think we're at an interesting point. So there are a few regulations that are emerging there are a few that are quite well developed and beginning to be to be implemented by regulators and by by regulated bodies um both within the eu but also in the the uk and also the us and singapore and other places um one of the one of the areas that's that's quite interesting i think particularly for for kind of connected data professionals is around taxonomies so you have a lot of regulators looking at the use of taxonomies in different ways to try and signpost green products and um

and divert um you know money that would traditionally go into kind of non-green or brown as they're sometimes called or blue kind of transition areas are diverted into into greener areas so the EU are really leading the way in that respect and have been first out of the blocks with the the green taxonomy there which is essentially classifying different industrial sectors and by virtue companies operating those those sectors as to whether they are considered green or not. Allied to that there's another there's another important regulation which is coming down the pipe, which is the SFDR, Sustainable Finance Disclosure Regulation, which is much, much more aimed at addressing things like greenwashing and looking at how financial products, particularly investment products, are labelled to consumers. So that's really exciting as well. So yeah, so I think over the next five years, we'll see a lot of kind of regulatory

activity and combined with this convergence of standards, I think the world of ESG will look very different by the time we get to the kind of 2030 timeframe. That sounds hopeful. And I really do hope that you're right. And there is some real progress made there as soon as possible, really. So at the moment, it looks like what companies like yours are doing are basically on a voluntary basis, I mean, from the side of your customers. So they're not really forced to comply to any sort of regulation, but I guess they do it for other reasons, to prepare themselves

for when the time comes or to reap other sorts of benefits like through social responsibility or marketing campaigns or whatnot. So maybe it's time that we go a little bit into more depth into the nature of what Neural Alpha does. So what are the services that you offer or the platforms that you're developing and who are your clients and what do you do for them? Sure.

So essentially, what does Neural Alpha do? Well, we are a data and software consultancy. So we predominantly build bespoke software and data products for financial institutions but also for ngos and civil society so kind of working outside in and trying to promote positive change through that way with the financial industry. As you probably guessed it with the nature of this podcast, we're huge connected data fans, huge graph fans, semantic technology, graphs, ontologies, taxonomies,

feature heavily in all of the work that we that we do um and i guess our particular sweet spot is applying those technologies to um esg issues that are typically obscured or hard to analyze but because of global supply chains and complex ownership structures. So we spend a great deal of our time looking particularly at things like deforestation, as an ESG thematic, particularly deforestation that's associated with soy and beef, palm oil and other soft commodities typically food based commodities although although some of the non-food based commodities things like metals minerals that kind of side of things um and the the i guess typically the the challenge with the deforestation and I guess why it's a particular focused area for us was prior to the last few years, it was very difficult to link on the ground deforestation happening in the tropics in places like Indonesia and the Amazon and other parts of the world to investors in, you know, New York or London or Singapore,

because of the reasons I mentioned in terms of, you know, there's many, many different actors involved in different parts of the supply chain. So yeah, so really, I guess this is a big, you know, it's a big sweet spot for graphs, being able to, you know, very rapidly do different traversals at a very low computational cost, being able to bring in different network analysis techniques, different graph analytics techniques. Let's try and pick up an example and see the different steps involved into doing what you do. So I guess the first step should probably be just getting hold of the right data sets. And I wonder what those data sets may be and where they may be coming from. Yeah, so we typically work with a wide variety of data sets and i guess that is another driver for why we use graph technology you know it's it's very very well equipped for doing data integration work particularly with messy data which is you know the the reality if you're working with particularly open data sources.

So we do a lot of work with things like trade data, import-export data that's come from bills of lading and other sources. We also integrate a lot of different tax database data and also ownership databases. and then try and match those to things like corporate registries to then disambiguate those and resolve them to common entities and start to be able to kind of link them into different data sets. Yeah, and then I guess another big kind of category of data that we do a lot of work with is financial data. So looking at things like risk factors, fundamentals, revenue costs data, looking at ownership data. So when we understand more about the different traders that are active, the different commodity traders that are active in different regions,

trading these forest risk commodities. Now, obviously, we then want to look a step removed at, you know, who are the direct and indirect investors that are exposed to those risks. So that typically involves bringing in a lot of financial data from, you know, people like Refinitiv and Frankset and Bloomberg and the like. So it sounds like you're basically doing a series of well a series of steps of association so for example like you have this area in Indonesia that you brought up recently and so who is managing that and then trying to figure out okay so who is. And so who is managing that and then trying to figure out, okay, so who is behind the entity who is managing that area? And so where are the

products going? And who owns, you know, the subsidiaries that are in touch with this intermediate entity, and so on and so forth, this sort of thing? Yeah, I mean, it's a's a massive over simplification but essentially in a nutshell um that that's kind of what we what we work through i mean on um our kind of real flagship project i guess is trace finance um which is focused around looking at these issues for these commodities fortunately we don't have to do all of this data integration and all of this modelling we tend to focus more on the finance part of the problem so fortunately we have really great partners in the form of Global Canopy

an NGO based in Oxford and also the Stockholm Environment Institute, who have really world leading sustainability researchers who really know these supply chains inside out. And that enables them to build kind of probabilistic models that will take a ton of export products and then be able to disaggregate that in it and assign it to different in-country infrastructure so in the case of soy for example you know you have things like soy crushing facilities and silos for storage in country and also at the ports. So they're able to then assign volumes to that infrastructure and then look at the region that then supplies that infrastructure and the deforestation that's occurring in that region

to then essentially calculate a deforestation exposure in hectares that is linked to a particular trader's, a particular commodity trader's sourcing practices. So fortunately, we have these really advanced models which we can build on when it comes to then looking at how those sustainability risks then translate in into the financial industry through you know different ownership structures different different lending structures yeah so it's yeah it's a it's a big challenge and yeah it's great to to play a part in kind of solving some of those problems. Okay, so actually, I wanted to ask you about the technology.

And you did mention the models that you help build. But before we get that part, I also wanted to ask about what you mentioned two of your most high profile projects for which I think you've also won some distinctions. So I wanted to ask, so right, your role in those projects is to basically provide the platform, let's say for those NGOs. So how are these NGOs going in turn to be using the platform? So do you know if they're going to be making that available to third parties such as researchers or journalists? What is the intended purpose there? Yeah, I mean, I guess mainly speaking about Trace and Trace Finance, which is the part

of the project which we lead. It's, sorry, what was the question was around? The overall scope, let's say, and goal of the project. Yeah, yeah. So in terms of that platform, it's, I guess, very much focused on providing tools for the financial industry to better understand how they're exposed to deforestation and to more proactively manage those risks within investment and loan portfolios. So although I guess we're developing these tools in conjunction with NGOs, and we have a lot of researchers and other civil society organizations that use these platforms.

I guess our main user base is actually the financial industry itself. very proactively using our data and our tools to do things like develop negative exclusion screens, positive inclusion screens, to perform monitoring, benchmarking of different holdings, different companies of interest. And I guess a big thing that we're spending a lot of time doing is also not just serving that to the industry as a software as a service solution, but also as a data product, and increasingly as a much richer commercial data products. Does that mean offering some sort of API so that people can consume it and integrate it in their own software stack or maybe also offering like data exports so again they can take their own data and utilize them in whatever way they wish? Yeah, essentially. So I guess the two main ways that the financial services industry, and indeed probably many

industries, consume this type of data is one through bulk data files. So things like delimited files tend to be one of the most common kind of paradigms that are that are followed and also for apis for more kind of integration in real-time requirements and that that's very interesting to us more for kind of like um uh point of sale or kind of real-time um requirements um so for example there's a lot of uh regulations of regulations that are currently being debated in the UK, and also within the EU and other parts of the world around supply chain due diligence, and particularly due diligence for banks that are wishing to lend to companies within the agricultural sector, where actually using things like real-time screening

is much, much more useful for those types of use cases. Okay, I see. So a bit earlier, you touched upon the actual technology that you use under the hood. Well, part of that I'm already familiar with, and there's a lot of, well, vocabularies and semantics and graph involved for the reasons that you mentioned. These technologies lend themselves well to use cases such as yours

in which you have lots of data from different sources to integrate. So that helps on the integration side. But you also mentioned the actual modeling that goes into that, which is part of the expertise that your NGO partners have in there. So I was wondering if you could shed a little bit of light on how those two play well together. Well, another point to mention there is that one of the benefits of utilizing graph technology is the fact that, again, besides the integration, you can basically do complicated, complex queries with lots of hops. And again, this is one of the requirements in your case. So if you can share a little bit more on the analytics side of things,

on what kind of investigations do you have to accommodate and how your technology helps do that? Sure. So I guess perhaps we're unusual or perhaps we're not very unusual in that I would say that we i would say that we have kind of two uh we have one foot in each of the kind of dominant camps i guess if you like within the kind of graph space so we have one foot within the the label property graph space and also another foot within more the kind of linked data rdf semantic web um space so we we do a lot of work um using label property

graphs for um different types of analytics so looking at kind of um you know shortest paths and community detection um computing things like influence um doing doing things with page rank and and this kind of side of the world um But we also do a lot of work, particularly when it comes to developing natural language processing models for things like topic modeling, concept extraction, entity recognition, relation extraction, using more kind of linked data standards. so particularly we do a lot of work building things like scoss taxonomies for breaking down you know thematics like like deforestation into kind of different related concepts that we're we're interested in

I guess yeah we also we also, I mean, that's touching a bit more on the graph side of things. We also do a lot of work with kind of more conventional NoSQL technologies, I guess. So we, you know, particularly do a lot of work with MongoDB, which I guess is pretty common these days. Most of our platforms are developed mainly using kind of JavaScript, Node.js, React, the kind of middleware front ends.

We do a lot of work with data visualization, D3. We also do a lot of work with graph visualization, particularly using things like Linkurious's Ogma framework, which we've found to be very performant in terms of particularly large-scale graph visualization. All right, so I guess we're pretty close to wrapping up. And so let's try and do that by asking you to sort of project where do you go from here, basically, not just for neural alpha, but where do you see the, how do you see this space evolving in general? I think for the last part of the question, you sort of touched upon that earlier. So convergence of standards and advancement of regulations,

which should help move the space forward. And how do you see Neuron Alpha evolving in this landscape as well? Yeah it's a really kind of thought-provoking question in terms of what the future holds. I wish I knew to be honest exactly, but I guess I don't know, I guess what do I hope that the future holds and where does that fit within some of the current trends that I see in terms of technology trends and also ESG scores and really trying to, you know, very, very much manage this, you know, this kind of inundation of data that people have by simplifying things in simple scores. And I think, you know, now a lot of people can really see the limitations of oversimplification that, you know, often in many cases, ESG scores are just not fit for purpose. So I think a lot of people are turning their attention more to actually using more efficient techniques and tools for, you know, being able to much more readily simulate and

integrate more of the raw data and actually really understand the context because i guess ultimately for me you know esg is it's an incredibly subjective space it's very context uh specific and that's the thing that personally i'm really excited about and excited about the direction that we're heading at in Neuron Alpha is how we can bring you know more kind of context rich tools to the market that you know enable people to really embrace this complexity and and and not not run away from it so um so you know in terms of what what what does that mean kind of on the ground I think um you know, much more use of graphs, much wider application of graphs and connected data technologies to other ESG topics. We're doing a lot of interesting work,

particularly around news processing and annotating news with different ESG concepts. I'm really excited about the direction that that could go in. And as part of that we're developing an internal ESG knowledge graph to try and apply some of these good practices to a much wider range of issues. Because it's kind of never-ending really within ESG. As science progresses, we're discovering new issues all the time out there and having to figure out how different companies are exposed to those issues. So yeah, I guess the genie is very much out of the bottle. And I guess we're always, on that note, I guess the genie is very much out of the bottle. And, yeah, and I guess we're always on that note, I guess, shameless kind of hiring plug,

but we're always looking for, you know, sustainability-centric kind of graph-literate engineers. So, you know, if these kinds of things excite people listening, then, yeah, we'd love to hear from you. I hope you enjoyed the podcast. If you like my work, you can follow Link Data Orchestration on Twitter, LinkedIn and Facebook.