Meeting Analytics & Deep Session Reports

plugNMeet's meeting analytics turns every session into a detailed analytics report: who joined and for how long, who spoke, who chatted, voted, raised a hand, shared a screen, or struggled with connection quality. Teams use it for meeting reports and attendance tracking; educators use it as a classroom performance report to measure participation equity, engagement trends, and technology barriers.

If enable_analytics was set in the room's room_features (room create API), plugNMeet aggregates analytics events in memory during the session and writes a single JSON file when the session ends. The file is stored as an artifact of type MEETING_ANALYTICS.

For the why behind these metrics — engagement questions, equitable participation, and privacy-by-default design — see From Black Box to Blueprint: Using Analytics to Understand Your Online Classroom.

What you can build from this data

• Meeting attendance report: per-user duration from joined / left timelines, plus room_total_users and room_duration.
• Participant engagement analytics: talked / talked_duration, mic_status, webcam_status, screen_share_status, raise_hand, voted_poll, public_chat / private_chat.
• Classroom performance report: combine talk time, poll votes, whiteboard annotation (whiteboard_annotated), chat files, and hand raises to compare participation across students or sessions. The Moodle plugin does this out of the box with cumulative attendance and activity completion rules — see Moodle integration.
• Connection quality report: connection_quality (excellent / good / poor) correlated with early left events to distinguish disengagement from network issues.
• Excel / LMS exports: the Moodle, WordPress, and Joomla plugins format the same data into per-room and per-user sheets. See WordPress artifacts for an example.

Analytics is opt-in and privacy-preserving: disabled by default, in-memory only, and permanently discarded if enable_analytics is false. Only metadata and counters are stored — never chat text, audio, or video content. See Security overview.

How to get the analytics file

1. Enable it at room creation: "enable_analytics": true in room_features.
2. Wait for the session to end. The server exports the JSON and emits the webhook event analytics_proceeded with the artifact_id — see Webhooks.
3. List files with /artifact/fetch using "type": "MEETING_ANALYTICS".
4. Get a one-time token from /artifact/getDownloadToken, then download: GET https://Your-Server/download/artifact/<TOKEN>

The deprecated /analytics/fetch and /analytics/getDownloadToken endpoints still work but will be removed. Use /artifact/* for new code.

Raw analytics JSON format

The raw file is intentionally low level — a complete event log, not a ready-made report. Don't aggregate it by hand; use the PHP SDK formatter below.

Top level is always:

{
  "room": {},
  "users": []
}

Room object

Base fields are written directly (see AnalyticsRoomInfo and exportAnalyticsToJSON in the server):

Field	Type	Notes
room_id	string	Your room_id.
room_title	string	Room title at end time.
room_total_users	int	Number of distinct users seen.
room_creation	int	Unix seconds.
room_ended	int	Unix seconds.
room_duration	int	Seconds, room_ended - room_creation.
enabled_e2ee	bool	From room metadata.
events	array	Low-level event list, see below.

User object

{
  "user_id": "111",
  "name": "Test User",
  "is_admin": false,
  "ex_user_id": "111",
  "events": []
}

user_id is the internal LiveKit participant ID, ex_user_id is the userId you passed in the join token (empty if auto_gen_user_id was used).

Event object

Both room and user events share the same shape:

{
  "name": "joined",
  "total": 6,
  "values": [{ "time": 1500000, "value": "1500000" }]
}

• name is the proto AnalyticsEvents name lowercased with the ANALYTICS_EVENT_ROOM_ / ANALYTICS_EVENT_USER_ prefix stripped, e.g. poll_added, recording_status, joined, mic_status. This stripping is done in processEventKey.
• time is Unix milliseconds.
• total / values encoding depends on how the server stored the event in Redis:
  • HSET (timestamped events like joined, mic_status): total = number of entries, values = full list.
  • INCRBY (counters like talked, public_chat): total = counter, values is empty.
  • SET (string values): total = 1, values has one entry.

Known event names (from plugnmeet_analytics.proto):

Room: poll_added, recording_status, rtmp_status, breakout_room, shared_notepad_status, external_display_link_status, external_media_player_status, speech_service_status, speech_service_total_usage, ingress_created, whiteboard_files, insights/transcription/translation counters.

User: joined, left, mic_status, talked, talked_duration (ms), webcam_status, screen_share_status, voted_poll, public_chat, private_chat, chat_files, whiteboard_annotated, whiteboard_files, speech_services_usage, speech_services_status, interface_visibility, raise_hand, connection_quality, reaction.

Status values inside mic_status / webcam_status / room status events use prefixes like ANALYTICS_STATUS_STARTED, ANALYTICS_STATUS_ENDED, ANALYTICS_STATUS_MUTED, or excellent / good / poor for connection quality.

Fields evolve with new features

New plugNMeet features can add or extend room fields, user fields, and event names at any time, so the lists above are a snapshot, not a fixed schema. Before parsing strictly or validating against hardcoded field names, always check the release notes and the authoritative plugnmeet_analytics.proto file. When using the PHP SDK, prefer getRoomFields() / getUserFields() over hardcoded lists so your code picks up new fields automatically.

Format with the PHP SDK

The plugNMeet-sdk-php package ships Mynaparrot\Plugnmeet\AnalyticsFormatter, which converts the raw JSON into an attendance-ready meeting report. The Moodle, WordPress, and Joomla plugins use the same class to generate their Excel exports.

composer require mynaparrot/plugnmeet-sdk

1. Fetch and download the JSON

<?php
require __DIR__ . '/vendor/autoload.php';

use Mynaparrot\Plugnmeet\PlugNmeet;
use Mynaparrot\PlugnmeetProto\FetchArtifactsReq;
use Mynaparrot\PlugnmeetProto\GetArtifactDownloadTokenReq;
use Mynaparrot\PlugnmeetProto\RoomArtifactType;

$pnm = new PlugNmeet('https://your-server', 'api-key', 'api-secret');

// list analytics artifacts for a room
$req = new FetchArtifactsReq();
$req->setRoomIds(['room01']);
$req->setType(RoomArtifactType::MEETING_ANALYTICS);
$res = $pnm->fetchArtifacts($req);

$artifactId = $res->getResult()->getArtifactsList()[0]->getArtifactId();

// download
$tokenReq = new GetArtifactDownloadTokenReq();
$tokenReq->setArtifactId($artifactId);
$token = $pnm->getArtifactDownloadToken($tokenReq)->getToken();

$json = file_get_contents('https://your-server/download/artifact/' . $token);
$raw = json_decode($json, true);

2. Format it into a report

use Mynaparrot\Plugnmeet\AnalyticsFormatter;

$formatter = new AnalyticsFormatter($raw, 'Asia/Dhaka');
$data = $formatter->getFormattedEventData();

$room  = $data['room'];   // associative array
$users = $data['users'];  // list of associative arrays

// helpers
echo $formatter->formatTimestamp($raw['room']['room_creation'], false);
echo $formatter->formatSecondsToTime($room['room_duration']);
print_r($formatter->getRoomFields());
print_r($formatter->getUserFields());

The wrapper in the SDK examples does the same in one call:

$formatter = plugNmeetConnect::getAnalyticsFormatter($json, $userTimezone);

Invalid timezones fall back to UTC.

What the formatter produces

Room fields (getRoomFields()):

room_id, room_title, room_creation, room_ended, room_duration,
room_total_users, enabled_e2ee, recording_status, rtmp_status,
shared_notepad_status, ingress_created, breakout_room,
external_display_link_status, external_media_player_status

Plus polls and whiteboard_files when present. room_creation / room_ended become d-m-Y H:i:s P strings in the given timezone. Status events become start-counts.

User fields (getUserFields()):

name, ex_user_id, is_admin, duration, joined, left, mic_status,
mic_muted, mic_duration, talked, talked_duration, webcam_status,
webcam_duration, raise_hand, voted_poll, whiteboard_annotated,
whiteboard_files, screen_share_status, public_chat, private_chat,
chat_files, interface_invisible, connection_quality

Key transformations:

• joined / left: sorted lists of Unix-ms timestamps.
• duration: total seconds present, computed from a join/left timeline. Duplicate joins are ignored, unmatched leaves are ignored, an unfinished session extends to room_ended, events outside [room_creation, room_ended] are ignored.
• mic_duration / webcam_duration: seconds between ANALYTICS_STATUS_STARTED and ANALYTICS_STATUS_ENDED (prefix match, so ANALYTICS_STATUS_STARTED:node_01 counts). An open device session closes at user left or room end. Duplicate starts are ignored.
• talked_duration: raw total is milliseconds → converted to seconds with ceil(ms / 1000).
• talked, public_chat, private_chat, chat_files, whiteboard_annotated, raise_hand: plain counters from total.
• interface_visibility → interface_invisible (count of hidden).
• connection_quality → { excellent, good, poor } counts.
• voted_poll: per-user vote count; also increments room.polls[poll_id].options[option_id].responses.
• duration, talked_duration, mic_duration, webcam_duration are clamped to min(room_duration, room_ended - room_creation). If room metadata is missing, the raw value is kept.

Formatted example — one row of a meeting analytics report:

{
  "room": {
    "room_id": "room01",
    "room_title": "Test Room",
    "room_creation": "17-05-2026 10:00:00 +06:00",
    "room_duration": 3000,
    "recording_status": 1
  },
  "users": [
    {
      "name": "Test User",
      "duration": 1200,
      "mic_duration": 500,
      "talked_duration": 82,
      "connection_quality": { "excellent": 10, "good": 2, "poor": 0 }
    }
  ]
}

Use formatSecondsToTime() to render duration fields as HH:MM:SS for tables or CSV/XLSX export.

Related resources

• Classroom analytics for educators — pedagogical guide to engagement analytics.
• Moodle integration — cumulative classroom performance reports with attendance and activity completion.
• Fetch artifacts API and download token API.
• Webhooks — analytics_proceeded event.
• Security overview — opt-in analytics and data retention.