Visual dashboards that show how your team communicates — message volumes, channel activity, speaker engagement, language usage, and time-of-day patterns.
The Analytics & Trends page provides a comprehensive visual overview of voice communication across your workspace. All data is computed from your existing message and channel data using pure SQL aggregations — no AI processing, no token costs, and no data leaves your database.
The dashboard is designed for administrators and team leads who want to understand communication patterns: Which channels are busiest? When is peak activity? Who are the most active speakers? What languages are being used?
At the top of the page you'll find two filter controls that apply to all dashboard panels:
After adjusting filters, click Refresh to reload all panels with the new parameters. Filters are preserved as you navigate between admin pages, so you can switch to other views and come back without losing your selection.
The top row shows six headline metrics for the selected time range:
These cards give you an at-a-glance snapshot before diving into the detailed charts below.
A time-series area chart showing message counts over time. The chart automatically chooses an appropriate time bucket:
Hover over any point to see the exact count and time. This chart helps you spot trends — is usage growing? Are there consistent slow periods? Did something unusual happen on a specific day?
A day-of-week × hour-of-day heatmap grid showing when your team is most active. Each cell is colour-coded by message intensity:
The heatmap is useful for workforce planning. If you see heavy activity on Monday mornings and Friday afternoons, you know those are peak periods. If late-night activity is unexpected, it might indicate on-call escalations worth investigating.
A horizontal bar chart showing the top channels by message count. Each bar represents a channel, sorted from most to least active. This quickly answers "Where is the most communication happening?"
Channels that appear unexpectedly high might be experiencing an incident. Channels with very low activity might be candidates for archiving.
A table ranking the top speakers by message count. For each speaker the table shows:
This is useful for identifying team members who are highly engaged — or for spotting situations where communication is concentrated in too few people.
A pie chart showing the breakdown of detected languages across transcripts. whoot.'s transcription engine supports 30+ languages, and this chart shows which ones your team actually uses.
For multilingual teams, this helps you understand language usage patterns. If a significant percentage of transcripts are in a language you didn't expect, it might indicate new team members or a regional shift in activity.
All analytics are computed by five optimised SQL functions that run directly in your Supabase database:
get_message_volume_trend — Time-bucketed message counts.get_channel_activity_stats — Per-channel message and speaker counts.get_speaker_leaderboard — Ranked speakers by message count.get_activity_heatmap — Day × hour aggregation grid.get_language_distribution — Language code grouping with counts.All five functions are called in parallel when you load or refresh the page, so results appear quickly even for large workspaces. There are no external API calls and no AI costs involved.
Viewing the Analytics & Trends dashboard requires the analytics.insights permission. This is the same permission that controls access to AI Insights. By default it is granted to the Admin and Owner roles.