Building a network of Impact Data Scientists

One of our primary goals at Kariba (the team behind Open Source Observer) is to build a network of Impact Data Scientists. However, “Impact Data Scientist” isn’t a career path that currently exists. It’s not even a job description that currently exists.

This post is our first step in trying to change that. In it, we discuss:

1. Why we think the Impact Data Scientist is an important job of the future

2. The characteristics and job spec of an Impact Data Scientist

3. Ways to get involved if you are an aspiring Impact Data Scientist

   Spoiler alert: join this groupchat and apply for data access here

One important caveat. This post is focused on building a network of Impact Data Scientists that serve crypto open source software ecosystems. In the long run, we hope to see Impact Data Scientists work in all sorts of domains. We are starting in crypto because there is already a strong culture around supporting open source software and decentralizing grantmaking decisions. We hope this culture of building in public and experimenting crosses over to non-crypto grantmaking ecosystems. When it does, we’d love to help build a network of Impact Data Scientists in those places too!

Why we think the Impact Data Scientist is an important job of the future

Crypto foundations and ecosystem funds are sitting on massive treasuries. Some are approaching nation-state levels of resources, measured in billions.

To make productive use of these assets, they award grants to open source developers who are building applications, libraries, and various forms of core infrastructure.

However, most foundations struggle to do grantmaking well. They have a hard time articulating what kind of impact they want to see more of and how that impact will be measured across their network. This is more than just a communications problem. It prevents foundations from effectively channeling their resources into the initiatives that are driving positive growth and adoption. Without good measurement of impact, there is a risk of wasted funds, underutilization of potential, and a lack of direction for builders.

For a case study, see our recent post on Optimism RetroPGF 3.

Much of the effort to date has gone into weeding out all the bad ways of spending money: the cash grabs, shell projects, sybil attacks, airdrop farming attempts, etc. There has also emerged a number of battle-tested projects that crypto foundations view as safe bets. This type of filtering work on both sides of the distribution curve is very important. Yet the strategy for every foundation can’t simply be to avoid the really bad projects and only fund the really good projects that everyone knows about already.

Instead, foundations should aspire to improve their grantmaking in a number of areas. They should get better at rewarding high impact but overlooked contributors. They should get better at spotting high potential projects on the long tail and accelerating their trajectory. They should get better at determining who are the keystone species in their ecosystems and what’s required to keep them healthy. They should get better at identifying the marginal value brought by each new initiative in relation to the ones that came before it. This is by no means an exhaustive list of improvement areas, just a starting point.

We believe the cumulative effect of such improvements can be an order of magnitude better than the status quo. However, to bring about this transformation, we need a new skill set to take root and flourish in the crypto grantmaking community. It’s the skill set of the Impact Data Scientist.

The characteristics and job spec of an Impact Data Scientist

Since there aren’t any examples of people who already have this job, let’s start by sketching out some of the characteristics of the people who might be intrinsically motivated to become an Impact Data Scientist:

🌱 You’re an optimist who is dissatisfied with the status quo. This dissatisfaction may stem from life experiences, things you’ve read, the environment you were born into, or just a vague sense that Moloch is all around us. Whatever the backstory, you are effective at channeling this energy into imagining positive outcomes. As an Impact Data Scientist, you wouldn’t yell and throw stones; you would look for ways to build things better.

🌟 Impact is your north star, but you are not beholden to a single definition of impact. You seek to accelerate impact in various forms. You acknowledge that to accelerate impact you need to measure it, and that measurement is fraught with problems. This complexity doesn’t demotivate you, however. You accept that impact measurement is like an ocean: from the shoreline it looks smooth and monotone, but once you start wading in you realize it contains entire worlds. This expansiveness motivates you to go deeper, to seek better and better explanations, and to dive well below the surface. When you discover things, you bring them back to the surface in the form of recommendations, practical insights, and new instruments. As an Impact Data Scientist, you would embrace complexity but you would also invite others to jump in and explore the waters with you.

💸 You are not a naive idealist. You know that funding matters. You know that to create more impact you need to reward at least some of it, financially. You want to create loops where value flows upstream in response to impact captured downstream. You believe in markets but you also see market failures all around you. You see markets as a game but decidedly not a zero-sum game. As an Impact Data Scientist, you would pursue the Platonic ideal of increasing the amount of positive impact per unit of money spent.

🏐 You are excited about designing better games. You are curious about how behavioral psychology, mechanism design, cognitive biases, and measurement instruments affect outcomes. Specifically, you want to apply these concepts to the allocation game. The allocation game is a game of allocating pools of money across different sets of projects based on impact. Even though you’ve never explicitly sat down to play the allocation game, you realize that you’ve been playing allocation games (designed by other people) your whole life. You play them when you vote, when you create a playlist on Spotify, when you leave money in the tip jar. You recognize that the result of these allocation games is as much a function of the rules of the game as it is the people invited to play. As an Impact Data Scientist, you would be hands-on crafting and refining some of these rules. You would want to use your unique blend of skills to enhance the fairness, effectiveness, and fun of these economic and social games.

✨ You know that data magic has an essential role to play in all this. Not just the numbers, but the stories and the visualizations that can be brought to life from a deep understanding of those numbers. In a digital world, data is the aggregated voice of its residents. You like quant stuff, and breathe a sigh of relief knowing that as an Impact Data Scientist you’ll have access to lots of clean data. Hold that thought. You have most likely been trained in an environment where data is served to you in long rectangles, where the rows/columns are fixed and the available feature set has been pre-specified. Are you ready to step out of the proverbial cave? If so, then there is a world around you that is feature-rich and pliable, full of things you can use to build your own rectangles (and other shapes). All you need is a single public key address to construct a social graph and an economic history. As an Impact Data Scientist, you would need to become a cartographer of this new data landscape.

If these possibilities capture your imagination, then read on… You appear to have many of the intrinsic motivations of an Impact Data Scientist!

What about the hard skills?

It is helpful but by no means essential that you are already proficient at coding and linear algebra. One of our goals at Kariba is to expand the opportunity space and create an attractor for people of different backgrounds and skill sets to explore these waters. So, bonus points if you’re already a Pythonista, but we also want to attract people who are good at design, writing, education, and governance to work on impact data problems.

What would be the job spec of an Impact Data Scientist? Here are some of the requirements we’d be looking for:

• Stack and connect the data. You love to discover, combine, and analyze novel datasets. You're excited about first connecting EVM data to off-chain code repositories, and then expanding to all sorts of domains where digital public goods interface with our everyday lives.
• Bring rigor to impact measurement. Imagine if all the analytical horsepower that goes into running a hedge fund or being number one in an e-commerce market were applied towards tracking open source impact? You’re ready to learn the dark arts so you can conjure them for the public good.
• Conduct hypothesis-driven inquiry. Embrace a scientific mindset, formulating hypotheses and critically evaluating biases without getting too lost in the theory. Your job is to seek better explanations for complex problems, but to be practical in testing the accuracy and reach of your explanations in the real world.
• Nerd out. Go down rabbitholes and apply empirical data to behavioral psychology, game theory, and human biases. Your understanding of these elements is crucial in shaping funding allocation outcomes and governance dynamics. You should find yourself in settings where you are an easy nerd snipe. You should hone the skill of nerd sniping others.
• Build in public. Champion open source principles, seek out criticism, engage in a robust peer review process across our digital commons and decentralized networks. Your voice is critical in spreading a culture of transparent, iterative learning and discovery. Ship early and check your ego at the door.
• Make an impact. Apply your skills in communities and causes that you care about and are affected by. Build credibility and show up consistently in those places. It’s very important to find practical use cases for your work. In our experience, this is often best achieved by serving people you already have some connection to and want to achieve their impact goals.

Again, we are NOT hiring right now, but we are looking for aspiring Impact Data Scientists to step out of the woodwork and join our community.

Ways to get involved if you are an aspiring Impact Data Scientist

Here’s the master plan:

1. Assemble a close-knit group of aspiring Impact Data Scientists and emerging practitioners
2. Build infrastructure, tooling, and educational content to serve their needs
3. Focus their analytical horsepower on ecosystems that are deploying significant capital to open source and that stand to benefit most from impact tracking
4. Reward contributors retroactively for the impact of their work
5. Keep iterating until Impact Data Scientist becomes an actual full-time role, either from within or as alumni of the collective

If this speaks to you, then there are two action items:

1. Join this groupchat. We are also working with other practitioners and theorists as a part of regenlearnings.xyz, helping with the “empirical data” vertex of the triangle.

2. Apply for data access here. Membership to the Kariba Data Collective is free but we want to keep the community close-knit and mission-aligned. Check out the contributing section of our docs for some specific examples of the infrastructure and tooling we’re looking to offer Impact Data Scientists.

Together, let's make 2024 Year 1 of the Impact Data Science movement!