Read results - Poll the Room

room_results answers with the session’s numbers, one snapshot per poll in deck order, and the same results as markdown.

The session

Field	Meaning
`title`	The deck’s name.
`status`	`lobby`, `presenting` or `ended`.
`joined`	How many people joined.
`pollCount`	Polls in the deck.
`avgResponseRate`	Answers over people joined, averaged over the polls, 0 to 1.
`liveMinutes`	How long the session ran.
`endedMinutesAgo`	When it ended, if it did.

Per poll

Every snapshot carries type, prompt, status (pending, live or final) and totalAnswers. The rest depends on how the widget takes answers.

Widgets	What the snapshot holds
Multiple choice, image choice, this or that, ranking	`options` with an `id` and `label`, and `counts` by option id. Ranking counts are Borda points.
100 points	`counts` are points spent per option.
Star rating, NPS, estimate	`average`, and `distribution` by bucket. Estimate adds `estimateAnswer` once final.
Radar	`dimensions` and `radar`, the room’s average per axis.
Word cloud, live map	`words` with `text` and `count`.
Open feed	`texts`, newest last.
Q&A	`qa` with `text` and `votes`, sorted by votes.
Spin the wheel	`winner`.
Place map	`pins`, one per person.

The markdown

markdown is the same results as a summary, one section per poll, ready to paste into a mail or a report:

# Town hall

Room 491026. 42 joined. 4 polls. Live for 38 minutes.

## Which one?

Multiple choice. 34 answered.

- A: 12 (35%)
- B: 18 (53%)
- C: 4 (12%)

An agent that reads the results can write the follow-up. That is usually the next thing the presenter asks it for.

Results are available while the room is held in memory on the server, which is until it restarts. Keeping them for good is on the roadmap.