• plantbo
  • Posts
  • Data A&R, But Make It Web3 šŸŖ™

Data A&R, But Make It Web3 šŸŖ™

Exploring Audius to find out whoā€™s building a fanbase in plain sight.

Before we dive in, I just want to give a quick shoutout to anyone who's recently joined this newsletter or has been sticking around for a while (even through the quieter months). šŸ˜„

And to anyone whoā€™s wondering what exactly Iā€™m trying to do here, I think this meme pretty much sums it up:

Honestly though, with AI, itā€™s even more feasible than ever.

Why isnā€™t anyone talking about Audius?
Seriously, we love to throw around terms like direct-to-fan, superfans (ugh, that word again), and fairer deals for artists. Yet somehow, Audius barely gets any love.

So hereā€™s why I think it deserves a bit more love:

  • Itā€™s helping to decentralize the music industry.

  • Fans can support artists beyond just pressing play.

  • It could be a goldmine for discovering forward-thinking artists who are tech-savvy and open to new models.

  • And hereā€™s a big one: platform governance. Unlike the popular DSPs where decisions just... happen, Audius lets token holders vote on changes. The power dynamics are different ā€” and thatā€™s kind of the whole point.

  • They are giving the artist, fan, and music industry way more valuable data points

The more I poked around Audius, the more I started thinking: if I were doing A&R at an indie label, how could I use this data? Maybe to:

  1. Get a better sense of who these so-called ā€œsuperfansā€ actually are.

  2. Spot artists in niche genres gaining traction and experimenting with new ways to build their careers.

Letā€™s dig in.

I pulled together a batch of 147 electronic artists whoā€™ve made a mark on Audius ā€” in other words, these are artists whoā€™ve already shown up on the platformā€™s charts. You could expand this pool, but right now, weā€™re just trying to prove a concept. After all, 147 artists x (potentially) 100 supporters x (potentially) 100 other artists they support... thatā€™s already a pretty solid discovery pool to explore.

To start, I wanted to answer two basic questions:

  1. Whoā€™s been getting the most $AUDIO support overall?

  2. Whoā€™s attracting the most distinct supporters (a.k.a., how many different people are backing them)?

Next up, I took those 147 electronic artists, grabbed up to 100 of their supporters each, and checked out which other top 100 artists (if they reached 100) those fans are also backing.

Iā€™ve always loved this kind of overlapā€”itā€™s one of the clearest signs that if someoneā€™s into Artist X, theyā€™re probably into Artist Y too. And on Audius, that signal is even stronger because we're not just talking about streams or follows ā€” these fans are actually putting $AUDIO behind the artists they care about.

Now imagine if you could track down specific supporters who are always early adopters, or even spot niche subgroups that keep showing up in different artist circles. That could be huge for music discovery.

The calculation behind this visualization measures the similarity between supporters by comparing the artists they each back, with closer supporters sharing more common artist preferences. This allows us to identify distinct subgroups of supporters who share similar tastes, and in the future, we could explore how early these groups are in supporting new or emerging artists.

Hold tight, weā€™re almost thereā€”just one last thing to explore!

I also wanted to see if a network graph could give us more insights by visualizing two tiers of support. Hereā€™s the breakdown:

šŸ§¬ 1. Seed: Your 147 Electronic Artists
These are your ā€œroot nodes,ā€ the starting point of the network.

šŸ™Œ 2. Supporters who backed them
Only the supporters who backed these artists with ā‰„1 support. These become the ā€œbridge nodesā€ that connect the artists.

šŸ” 3. The artists the supporters from the seed artists also backed
Now we get to see which artists share fan bases with the original group and what new scenes or sounds these fans are exploring.

This gives us a clearer picture of which artists are connected through their shared supporters and could help us spot emerging trends, crucial supporters or discover new artists.

artists with a lower # of connections were filtered out

  • Black represents the seed artists (the starting point of the analysis)

  • Green shows the first-tier supporters, those fans backing the seed artists

  • Red represents the artists supported by the first-tier supporters, revealing further artist connections

I couldā€™ve spent more time refining this, as a data visualization fan šŸ˜¬, but the goal here is really to ship ideas and theories.

When certain supporters and artists appear together in the plot, it suggests that these fans are backing multiple artists within the same cluster, hinting at shared musical tastes or overlapping fan bases.

I realize Audius has a smaller user base compared to the bigger DSPs, but the way it's set upā€”decentralized and with so much public data that other platforms donā€™t haveā€”opens up some really cool ways to discover new talent.

That's a wrap for this week! Let me know what you think by voting in the poll below. Catch you next time! šŸ‘‹

Did you enjoy this newsletter?

Login or Subscribe to participate in polls.

Iā€™ve been wanting to share more of these little scripts Iā€™ve writtenā€”mainly because Iā€™m all about making data (especially from public sources or content creators) easier to understand. Audius does a pretty solid job of this, but letā€™s be realā€¦ some of the other DSPs could definitely learn a thing or two. šŸ˜…

I know this might be more up the alley of the data nerds out there, but I also know this newsletter has a bit of a mixed crowd. So hereā€™s a Python script that grabs genre chart data from Audius over different time periods. Enjoy!

Reply

or to participate.