Data Collection

Liwan is self-hosted web analytics. Events are processed by your Liwan server and stored in your deployment, not on Liwan-managed cloud servers.

Liwan is designed to avoid cookies, cross-site tracking, and long-lived visitor identifiers. It still collects analytics events, so you should choose collection settings that fit your site and legal requirements.

Recommended Defaults

If you want a more conservative setup:

Use Random per request or Network standard visitor grouping.
Use Country only or None for geolocation.
Disable UTM parameters unless you need campaign attribution.
Disable Session metrics unless you need bounce rate, time on site, entry page, and exit page.
Use a retention period around 6 months.
Add drop rules for internal, local, test, or other unwanted traffic.

Where Data Is Processed

Liwan runs on your server or hosting environment. The browser sends tracking events to your Liwan instance, and your instance stores the resulting analytics rows.

The raw IP address is used during request processing for visitor grouping and optional geolocation, then discarded. It is not stored in the events database.

What The Tracker Sends

The browser-side tracker can send:

Page URL: the current page URL, with URL fragments removed and only attribution query parameters preserved.
Referrer: where the visitor came from, if the browser provides it.
Attribution query parameters: campaign-style query parameters in the page URL. Liwan stores them separately only when UTM tracking is enabled.
Screen width bucket: a coarse bucket such as SM, MD, or XL.
Orientation: portrait or landscape.
Event name: currently pageview-style events unless you use custom tracking code.

Liwan strips query parameters and URL fragments from stored page URLs. The default browser tracker also ignores localhost, loopback IPs, and file:// pages.

What The Server Derives

The Liwan server can derive additional fields from the request:

Visitor grouping ID from the request IP address, user agent, entity ID, and daily salt, depending on the selected visitor grouping mode.
Browser and platform families from the user agent string, without full version strings.
Device type from the user agent parser.
Country and city from GeoIP lookup, if geolocation is enabled and configured.
Session intervals for bounce rate, time on site, entry page, and exit page, if session metrics are enabled.

Liwan also drops obvious bot traffic and spam or local referrers before storage.

Visitor Grouping

Visitor grouping controls how repeat visits are counted without storing raw IP addresses.

Mode	What it does	Privacy note
Accurate	Hashes IP address, user agent, daily salt, and entity ID.	Groups visitors most accurately, but uses hashed fingerprinting (IP + User Agent) during processing. Still rotates daily and does not store the raw inputs.
Random per request	Generates a random visitor ID for each event.	Minimizes repeat-visitor recognition because events are not linked by IP address or user agent.
Network standard	Hashes a masked network prefix (`/24` IPv4, `/56` IPv6) with daily salt and entity ID.	Avoids exact-IP grouping and does not use user agent in the grouping hash.
Network balanced	Hashes a narrower masked network prefix (`/28` IPv4, `/64` IPv6).	Uses a more specific network prefix than standard mode, but still avoids user agent in the grouping hash.
Network accurate	Hashes the full IP address with daily salt and entity ID, without user agent.	Uses exact IP during processing, but avoids user agent in the grouping hash and rotates daily.

The daily salt rotates, so visitor grouping is not persistent across days.

For a more privacy-oriented setup, use Random per request or Network standard. IP-based network modes group traffic by network address blocks, not individual humans.

Geolocation

Geolocation can be disabled or limited:

None: do not store country or city.
Country: store only country.
Country and city: store both country and city.

GeoIP lookup uses the request IP during processing. The IP address itself is discarded after processing.

If you do not need city-level metrics, use Country only or None.

UTM Parameters

When UTM tracking is enabled, Liwan stores these campaign fields from page URLs:

utm_source
utm_medium
utm_campaign
utm_content
utm_term

Liwan also accepts common aliases such as source, medium, campaign, content, term, ref, referrer, and referer, and maps them into the same stored UTM fields.

Disable UTM tracking if you do not need campaign attribution or want to reduce stored marketing data.

Session Metrics

Session metrics are required for:

Bounce rate
Average time on site
Entry page
Exit page

When enabled, Liwan stores intervals between pageviews for a visitor group. When disabled, those dashboard metrics are unavailable or less complete.

Drop Rules

Drop rules remove matching events before they are stored.

Each rule is checked separately. Inside one rule, all filters must match. Matching any rule drops the event.

Examples:

Drop local development traffic by domain or path.
Drop internal pages.
Drop traffic with a known test campaign value.
Drop events from a specific country or city if geolocation is enabled.

Global drop rules apply to every entity. Entity drop rules are additional rules for that entity only.

Retention And Pruning

Retention controls how long historical events are kept. You can keep all history or automatically prune events older than a selected number of days.

The pruning tool can also apply current collection settings to historical data:

Clear UTM fields when UTM tracking is disabled.
Clear geolocation fields when geolocation is disabled or reduced.
Clear session interval fields when session metrics are disabled.
Delete events older than the selected retention period.

Run a dry run first to preview how many events would be affected.

A 6 month retention period is a reasonable default for many sites.

Operational Notes

Liwan uses a daily rotating salt and does not store raw IP addresses. In practice, Liwan does not keep enough information to look up one specific visitor’s history across multiple days, which is a consequence of its data-minimization approach.

Because Liwan is self-hosted, your analytics data stays in your infrastructure rather than being sent to a third-party analytics vendor.

What Liwan Does Not Store

Liwan does not store:

Raw IP addresses.
Cookies for visitor tracking.
Cross-site visitor identifiers.
Full referrer paths or query strings from normal URL referrers.
Page URL query strings or fragments, except extracted UTM parameters when enabled.
Visitor identifiers that persist across multiple days.

Liwan still processes request data such as IP address and user agent to produce analytics. If you need stricter behavior, use the collection settings above, add drop rules, or change the tracker or server behavior.