Reading the statistics page
The Stats page is the dashboard view you'll probably visit most often. It packs your server's moderation activity into roughly a screenful — KPIs at the top, a time-series chart, breakdowns by category, action, and language, top tables, and a real-time feed.
Opening Stats#
Open Stats from the dashboard sidebar, then pick the server you want statistics for (or open Stats from inside a specific server). The page is per-server — to compare servers you'd open them in separate tabs.
Filters#
The filters above the charts:
- Date range. Presets for today, last 7 days, last 30 days, last 90 days, or a custom range.
- Category. Narrow to one of the six moderation categories (Toxicity, Harassment, Spam, NSFW, Hate speech, Self-harm).
- Action. Narrow to a specific action (none, flag, delete, timeout, ban).
Filters apply to all sections of the page simultaneously, including the CSV export.
KPIs#
A strip of headline numbers across the top: how many messages were processed, how many were flagged, how many were deleted, how many timeouts and bans were applied. The numbers reflect the current filter selection.
Time-series chart#
A chart showing moderation activity over time, with the bucket size automatically chosen to match your date range (hourly buckets for "today", daily buckets for longer ranges).
Patterns worth watching:
- A spike on a specific day. Often a raid, a controversial topic, or a recent threshold change.
- A steady upward trend. Either the community is changing, the thresholds are wrong, or both.
- Big flag rate, small delete rate. You're putting a lot of borderline cases in the review queue. Either tune thresholds tighter or relax them.
Category breakdown#
A donut chart showing which of the six categories — Toxicity, Harassment, Spam, NSFW, Hate speech, Self-harm — are driving moderation activity. A big slice for a category that's normally small (e.g., a sudden spike in hate speech) is worth investigating in the audit log.
Action breakdown#
A bar chart showing what action was taken across the date range: none, flag, delete, timeout, ban.
Language breakdown#
For multilingual servers: a chart showing which languages Civora detected in the moderated content. Useful for spotting spam or raid activity — automated bots often arrive in one source language regardless of where they're targeting.
Top channels / Top users#
Two tables:
- Top channels by moderation activity. Channels with the most actions taken. If one channel dominates, decide whether that's expected (a raucous general chat) or something to fix (rules that don't fit that channel — exclude it or tune it).
- Top users. Users with the most actioned messages. Usernames are obscured by default for privacy (shown as
kova***_p). Hover or focus a name to reveal the full username in a tooltip — there's a GDPR notice above the table explaining the reveal-on-hover behavior.
Live feed#
A column streaming new moderation events as they happen in real time. Useful during a known-stressful event, to confirm a threshold change took effect, or just for a quick sanity check that Civora is still doing reasonable things. There's no refresh button — the feed updates itself.
CSV export#
An export button on the Stats page lets you download the raw event log for the current filter selection as a CSV. Useful for analysis in a spreadsheet, sharing data with a stakeholder who doesn't have dashboard access, or feeding it into your own analytics pipeline.
What's next#
- Audit log — the per-message historical record. Stats summarizes; the audit log lets you investigate.
- Severity tuning — the lever that explains most of what you'll see on the Stats page.