How do you analyze community and events data in Metabase?
Community and events tools hold the record of who shows up: every member, post, thread, RSVP, and check-in. In Metabase, land that record in a SQL warehouse and model layers for members, activity, events, and registrations. Those layers support dashboards for member growth, engagement, response times, registrations, and attendance — next to the product, support, and revenue data the rest of the company already reports on.
TL;DR — This is the least connector-friendly category on the site. Only Eventbrite has a well-covered managed connector; Circle has none at all. Plan on scheduled API pulls or a database replica, define "active member" once and state its window everywhere, and never let registrations stand in for attendance.
Which tools does this cover?
- Discourse — topics and posts, users and trust levels, categories and tags, likes and reactions
- Circle — members and posts, spaces and space groups, comments and reactions, events and RSVPs
- Discord — messages, guild members and roles, channels and threads, reactions
- Luma — event guests, events and calendars, ticket types, check-ins
- Eventbrite — attendees, events, orders and ticket classes, venues
Two disambiguations, because both names are heavily overloaded: Circle here is circle.so, the community platform — not Circle Internet Financial, the USDC company. Luma here is lu.ma, the events platform — not Luma AI, the video and image generation company.
For the internal-collaboration side of the same story, see Slack; for the ticket-driven side, see the customer support category, where the same people show up with a different label on them.
What is the shared community and events data model?
| Concept | Common terms | Used for |
|---|---|---|
| Member | User, member, guild member, guest | Growth, activation, retention, segmentation |
| Post or message | Post, comment, reply, message | Engagement volume and contributor concentration |
| Topic or thread | Topic, thread, discussion | Response times, unanswered questions |
| Space or channel | Category, space, channel, forum | Where activity concentrates, and where it stopped |
| Event | Event, meetup, session, scheduled event | Programming performance and capacity |
| Registration | RSVP, guest, attendee, ticket | Signup volume, pacing, and ticket mix |
| Check-in | Check-in, scan, attendance | The only real evidence anyone attended |
| Order | Order, transaction, refund | Ticket revenue net of fees and refunds |
How do you connect community tools to Metabase?
- Database replica route — the best case, and available for self-hosted Discourse: replicate the forum's PostgreSQL database into your warehouse and every table is yours.
- Pipeline route — a managed connector where one exists (Eventbrite, Discord, Luma) or scheduled API pulls where none does (Circle), landing raw tables in a warehouse.
- MCP or API + CLI route — pull a scoped, summarized slice through a vendor MCP server or a scripted API export, save CSV, and load it with
mb upload csvfor fast exploration.
A note on MCP in this category, since the marketing is noisy: Discourse publishes its own server, and Circle hosts one for admins on Business plans and above. Discord ships an official MCP server for its developer documentation only — it exposes no guild, member, or message data, so community analytics still runs through a bot. Luma and Eventbrite publish no MCP server at all (verified July 2026); their guides lead with a scripted API export instead.
What can you analyze across these tools?
- Member growth rate— net new members per period, from daily snapshots rather than the live list
- Engagement rate — members who actually did something, over a window you state on the card
- RSVP-to-attendance conversion — check-ins over registrations, for events where check-in was used
- Reply rate — the share of questions that get an answer, and how long it takes
- Conversion rate — activation of new members, and free-to-paid membership where that exists
- Contributor concentration — the share of all posts written by the top ten members, the number nobody wants to look at
- Ticket revenue net of refunds — gross, fees, and refunds kept as separate columns, never collapsed into one
Which dashboards should you build?
- Member growth — new members, net growth, and activation of recent signups
- Engagement activity — posts, messages, and reactions by space, channel, and month
- Community health — response times, unanswered questions, and contributor concentration
- Event performance — registrations, attendance, ticket mix, and revenue by event
Common mistakes
Reporting active members without defining active.→ Thirty days or ninety, posted or merely logged in — each choice gives a different number and none is a standard. State the window and the action on every card.
Recomputing member history from the current member list.→ Deleted and anonymized accounts vanish from it, so last quarter's growth changes every refresh. Snapshot member counts daily into a table you own.
Treating registrations as attendance.→ Attendance is unknowable without check-in data. Exclude events where nobody checked anyone in, and show how many you excluded.
Assuming a managed connector exists.→ Several of these platforms have none, and some that do ship full-refresh streams only. Check the connector's streams and sync modes before you plan a dashboard around them.
Counting membership as community health.→ A big member list with ten people posting is not a community. Pair every growth number with an engagement number and contributor concentration.
Related
Analytics
Integrations
Metrics
FAQ
What is community analytics?
Community analytics is reporting built on the raw records of a community — members, posts, topics, channels, events, and registrations — modeled in a SQL warehouse and analyzed in Metabase. The built-in dashboards inside a forum or community platform answer "how are we doing this week?"; community analytics in your own BI answers "does community participation change retention, support load, and revenue?" See the community analytics overview for the full build.
Does Metabase connect natively to Discourse, Circle, or Discord?
No — and no BI tool honestly does. Metabase reads SQL databases and warehouses, so community data has to land in one first. Discourse is the easy case if you self-host: the forum runs on PostgreSQL you can replicate. Circle and Discord both require a pipeline you build — Admin API pulls for Circle, a bot-token connector for Discord. Every guide also documents the quick route: export a scoped CSV and load it with
mb upload csv.Which of these tools have managed connectors?
Fewer than you would hope. Eventbrite is the best served — an Airbyte Marketplace source with incremental support on most streams, plus a Fivetran Application Lite connector. Discord has an Airbyte Marketplace source whose messages stream is full-refresh only. Luma has a two-stream Airbyte source, also full-refresh. Discourse has a Fivetran Lite connector and no Airbyte source. Circle has nothing at all: schedule Admin API v2 pulls yourself.
How do I measure event attendance?
Only from check-in data. If nobody scanned tickets or ran check-in at the door,
checked_in_at is null for reasons unrelated to who actually showed up — so those events cannot be scored. Exclude them from RSVP-to-attendance conversion and count them on the dashboard instead. Registrations are never a stand-in for attendance.Which tools does the community and events category cover?
Five guides: Discourse for forums, Circle for hosted community platforms, Discord for community chat, and Luma and Eventbrite for events. All five map onto a shared model — members, posts, topics, spaces or channels, events, registrations, and orders.
Which dashboards should you build first?
Start with member growth — it needs only a member table with join dates and a daily snapshot. Add engagement activity once posts or messages are landed, community health when you can compute reply times, and event performance as soon as registrations and check-ins are in the warehouse.