Data / Ecosystem

Ecosystem data

Daily series describing what a project actually does when the price is not moving: how its audience grows, how much code lands, and how many people write it.

69
assets with ecosystem series
curated, not the full registry
5
independent daily series per asset
daily
collection cadence
fixed time per series
2024
earliest retained records
varies by asset and series

What this data is

Market and social data describe attention. Ecosystem data describes activity. The two diverge constantly, and that divergence is most of the value here: a project whose contributor count has been flat for a year while its social volume triples is telling you something that neither series says on its own.

Coverage is a deliberately curated set of 69 established assets rather than the full 7,461-asset registry. Ecosystem series require a mapped X account, community source and repository per project, and that mapping is maintained by hand. We would rather cover 69 assets correctly than several thousand badly.

The five series

SeriesCadenceFieldsWhat it measures
Channel reachdailyfollowers, following, tweetsAudience size of the project's primary X account and how much it posts.
Communitydailysubscribers, activeUsers, likes, sharesSubreddit membership and activity. Frequently null where no community source is mapped.
Repositorydailystars, forks, subscribers, total_issues, closed_issues, pull_requests_merged, pull_request_contributors, commit_count_4_weeks, code_additions_4_weeks, code_deletions_4_weeksPublic code activity on the project's primary repositories.
Contributorsdailycontributors, contributionsDistinct contributors and cumulative contributions.
Topic mentionsdailydate, mentionsDaily mention counts for the project's topic, as a slower companion to the hourly social series.

Where it comes from

Each series is collected from the project's own public surfaces, resolved through the metadata provider's record for that asset: the X handle, the subreddit URL and the repository URLs. Those three pointers are the join key, and they are the part that most often needs correcting by hand — which is why the underlying records carry explicit override flags for each of them.

Override flags are part of the data

An asset record carries override_twitter_screen_name, override_subreddit_url and override_github_url. When one is set, the pointer was corrected manually because the provider's value pointed at the wrong account, a dead subreddit, or a mirror repository. Ecosystem series are only as good as those pointers, and the flags are how we keep track of which ones have been checked.

How it is processed

Asset record
canonical assetx handlesubreddit urlrepository urls
Daily collectionone fixed time per series
channel countsrepository metricscontributor countstopic mentions
Normalization
one row per asset per dayUTC timestampsnulls preserved as unknown
Series
append onlyjoinable to market and social series

Update frequency

Daily, with each series collected at its own fixed time of day so that collection load is spread and each series has a stable, comparable sampling hour. Measured collection times are listed on the freshness page.

Daily cadence is a property of the underlying sources, not a limitation we chose. Repository statistics and follower counts do not meaningfully change hour to hour; sampling them hourly would add rows without adding information.

Who this is for

Ecosystem series are the part of the layer that funds, analysts and project teams reach for most, because they are the part that is not already priced in everywhere. Typical uses: tracking whether a treasury-funded roadmap is producing commits, comparing contributor breadth across competing chains, or checking whether community growth preceded or followed a price move.

Limitations

  • Curated coverage. 69 assets, not the full registry. If an asset is not in the set, it has no ecosystem series at all.
  • Sparse community fields. Reddit figures are null for many assets, including large ones. Null is unknown, never zero.
  • Repository choice is a judgement. Most projects have many repositories. The series follows the primary ones recorded on the asset, so work that happens elsewhere is invisible to it.
  • Counts are cumulative, and that hides things. Stars, forks and contributions only ever rise. Read the rate of change, not the level.
  • Some fields are frequently unavailable upstream. Code addition and deletion counts in particular are often null even where commit counts are present.

Which applications use it

Ecosystem series drive the ecosystem activity and ecosystem score surfaces on moonboard.ai, where they are shown as trends over a selectable window rather than as point-in-time values.

Related