> ## Documentation Index
> Fetch the complete documentation index at: https://social-b97141fb-feat-user-history-moderation-detail.mintlify.site/llms.txt
> Use this file to discover all available pages before exploring further.

# Chat Analytics

> Operational insight into messaging volume, participation, channel health, quality, and moderation load to optimize chat retention and scalability.

<Info>
  Use Chat Analytics to detect participation shifts, capacity risks, emerging spam, and moderation workload so you can sustain healthy real‑time conversations.
</Info>

<CardGroup cols={3}>
  <Card title="Message Volume" icon="chart-column">Total & trend</Card>
  <Card title="Participation" icon="users">Unique senders & activation</Card>
  <Card title="Channel Health" icon="hashtag">Distribution & concentration</Card>
  <Card title="Engagement Quality" icon="face-smile">Reactions & replies depth</Card>
  <Card title="Moderation Load" icon="shield">Flagged rate & review pressure</Card>
  <Card title="Temporal Patterns" icon="clock">Daily & hourly peaks</Card>
</CardGroup>

## Fast Questions You Can Answer

<Tabs>
  <Tab title="Is growth broad?">Are more unique users sending messages vs just power users ramping?</Tab>
  <Tab title="Quality dip?">Are comments/replies (thread depth) or reactions per message falling?</Tab>
  <Tab title="Channel sprawl?">Are new channels fragmenting engagement (rising concentration index)?</Tab>
  <Tab title="Safety load?">Is flagged rate or flagged volume per moderator escalating?</Tab>
  <Tab title="Capacity risk?">Are peak concurrent message bursts tightening vs infra headroom?</Tab>
  <Tab title="Activation health?">Are newly activated senders retaining after day 1 / week 1?</Tab>
</Tabs>

## Daily Operational Workflow

<Steps>
  <Step title="Apply Filters">Pick date range (e.g., last 7 / 30 days), segment (environment, region), and exclude internal test channels.</Step>
  <Step title="Scan KPI Cards">Messages, Flagged Messages, Reactions, Unique Senders—compare % delta vs prior period.</Step>
  <Step title="Review Volume Trend">Open New Messages by Day to detect bursts or weekend/weekday abnormalities.</Step>
  <Step title="Check Participation Breadth">Contrast Users by Day with Messages—if messages up but users flat, investigate power user over-reliance.</Step>
  <Step title="Channel Distribution">Check Channels Count by Type and Messages by Channel for concentration or underutilized types.</Step>
  <Step title="Moderation Pressure">Flagged Messages & Flagged Rate vs moderation SLA; queue backlog may require staffing shift.</Step>
  <Step title="Temporal Hotspots">Use Heatmap to validate moderator coverage during peak hours.</Step>
  <Step title="Drill Power / Risk Users">Use Users table to inspect top senders & high flagged contributors.</Step>
  <Step title="Log Actions">Document interventions (rate limits, highlights, education) and set follow-up date.</Step>
</Steps>

## Modules & Interpretation

### KPI Cards

High-level snapshot (Messages, Flagged Messages, Reactions, Unique Senders). Deltas contextualize growth vs prior period.

### New Messages by Day

Volume rhythm; use for feature launch impact and anomaly detection. Gentle cyclic pattern is normal; abrupt plateau indicates engagement stall.

### Channels Count by Type

Composition (public, private, broadcast). Skew heavily toward one type may limit discovery or create moderation blind spots.

### Users by Day

Active unique senders; rising volume without matching unique sender growth implies intensity, not breadth.

### Total Messages by Type / Channel Distribution

Identifies reliance on a small set of channels. Over-concentration risks single-point community health issues.

### Top Group Chat Channels Leaderboard

Sort by Members, Messages, Engagement Rate (messages per member), or Flagged Rate to surface exemplars or risk clusters.

### Messages Heatmap

Hourly/daily density. Align moderator shifts and system scaling policies with dark (peak) cells.

## Key Metrics & Formulas

| Metric | Definition | Formula (Illustrative) | Why It Matters | Action Trigger |
| - | - | - | - | - |
| Messages | Total messages sent | count(messages) | Overall volume | Drop >15% w/w |
| Unique Senders | Distinct users sending ≥1 message | distinct(user\_id) | Participation breadth | Flat while messages rise |
| Messages per Active User (MPAU) | Avg intensity per sender | Messages / Unique Senders | Detect over-reliance | Sharp rise + flat senders |
| Reactions | Total reactions on messages | count(reactions) | Lightweight engagement | Reactions/Message down |
| Reaction Rate | Reactions per message | Reactions / Messages | Content resonance | Below baseline band |
| Flagged Messages | Messages flagged by systems/users | count(flag events) | Moderation workload | Spike >25% w/w |
| Flagged Rate | Flagged / Messages | Flagged / Messages | Safety cleanliness | > threshold (e.g., 1–2%) |
| New Channels | Channels created period | count(channel\_create) | Organic growth / fragmentation | Spike + falling engagement rate |
| Channel Concentration Index | Engagement distribution (simplified HHI) | Σ (channel\_share^2) | Diversity of conversation | > prior period + rising churn |
| Messages per Channel | Avg messages per active channel | Messages / active\_channels | Capacity & sprawl measure | Decline + rising channel count |
| Peak Hour Factor | Peak hour messages vs avg hour | peak\_hour / avg\_hour | Staffing & infra planning | Factor rising > target |
| Newly Activated Chat Users | First-time senders in period | distinct(new\_sender\_ids) | Feature adoption | Activation drop consecutive weeks |
| Retained New Senders (7d) | New senders returning within 7d | returning\_new / new\_sender\_ids | Early retention | \< benchmark band |

## Participation & Quality Lenses

<Tabs>
  <Tab title="Acquisition">Growth in newly activated chat users.</Tab>
  <Tab title="Activation">First 7d retention & reactions per new user's messages.</Tab>
  <Tab title="Retention">Unique senders cohort returning week over week.</Tab>
  <Tab title="Quality">Reaction Rate & flagged inversely correlated.</Tab>
  <Tab title="Safety">Flagged rate & backlog aging time.</Tab>
  <Tab title="Load & Capacity">Peak Hour Factor vs infra headroom.</Tab>
</Tabs>

## Benchmark & Governance Strategy

<AccordionGroup>
  <Accordion title="Baseline Collection">Capture 30 rolling days after stable launch; derive median & IQR for each core metric.</Accordion>
  <Accordion title="Adaptive Thresholds">Alert if metric leaves median ±1.5 \* IQR (robust vs outliers).</Accordion>
  <Accordion title="Cohort Drift">Track new vs established communities; avoid mixing maturity profiles.</Accordion>
  <Accordion title="Quality Gate">Require Reaction Rate & Retained New Senders above thresholds before promoting channels.</Accordion>
  <Accordion title="Moderation SLA">Define max backlog aging (e.g., 2h). Breach triggers escalation & auto-priority ordering.</Accordion>
</AccordionGroup>

## Early Warning Signals

<Tabs>
  <Tab title="Power User Dependence">MPAU rises, Unique Senders flat.</Tab>
  <Tab title="Fragmentation">Channel count up; messages/channel & engagement drop.</Tab>
  <Tab title="Spam / Flood">Messages spike; Reaction Rate collapses; Flagged Rate rises.</Tab>
  <Tab title="Moderation Overload">Flagged per moderator exceeding SLA; backlog aging up.</Tab>
  <Tab title="Quality Decline">Gradual fall in reactions & retention without volume change.</Tab>
  <Tab title="Adoption Stall">Newly activated chat users trending downward 2+ weeks.</Tab>
</Tabs>

## Troubleshooting

| Symptom | Likely Cause | Investigation | Resolution |
| - | - | - | - |
| Volume up; engagement flat | Few users posting heavily | Distribution of messages per user | Encourage broader participation; prompts, onboarding nudge |
| Flagged rate surge | Spam / abusive burst | Review top flagged channels & user IDs | Apply rate limits; tighten filters; temp-ban offenders |
| Reaction rate drop | Content low relevance / UI friction | Compare reaction latency post message | Improve quick reaction UI; highlight engaging threads |
| New channels spike; avg messages/channel down | Fragmentation / channel creation sprawl | Correlate channel age vs activity | Introduce channel creation guidelines & approval |
| Moderation backlog aging | Understaffed peak or automation gap | Backlog aging metric by hour | Add on-call; refine auto-flag thresholds |
| Low new sender retention | Poor onboarding into existing channels | Funnel: Signup → First message → Return visit | Provide channel suggestions & starter messages |
| Peak hour factor rising | Emerging synchronized usage cluster | Compare heatmap vs previous weeks | Pre-scale infra; adjust sharding or throughput configs |

## Related

<CardGroup cols={2}>
  <Card title="Activity Analytics" icon="wave-pulse" href="./activity-analytics">User activity baseline</Card>
  <Card title="Social Analytics" icon="chart-network" href="./social-analytics">Community performance</Card>
  <Card title="AI Social Insights" icon="sparkles" href="./social-insights">Qualitative topics & sentiment</Card>
  <Card title="Raw Data Export" icon="download" href="./raw-data-export">Custom message logs</Card>
</CardGroup>

<Info>Need custom spam detection dashboards or extended retention windows? Contact support for advanced analytics enablement.</Info>


This documentation is built and hosted on [Mintlify](https://mintlify.com), a developer documentation platform.