fix(analytics): use rolling 7-day window for WAU calculation by saddlepaddle · Pull Request #490 · superset-sh/superset

saddlepaddle · 2025-12-23T18:00:38Z

Summary

Fix WAU calculation to use a proper rolling 7-day window per day
Previously used non-overlapping weekly buckets which didn't show day-over-day changes
Now returns daily data points, each showing users active 3+ days in the preceding 7 days

Before vs After

Before: Weekly buckets (e.g., Week 1: days 0-7, Week 2: days 8-14)
After: Rolling window per day (e.g., Dec 23: days 17-23, Dec 22: days 16-22)

Definition

WAU = Users who had workspace_created events on 3+ distinct days within a rolling 7-day window.

Test plan

Verify WAU chart shows daily data points
Confirm numbers reflect rolling 7-day windows

🤖 Generated with Claude Code

Summary by CodeRabbit

Bug Fixes
- Updated WAU Trend Chart to use rolling 7-day calculation methodology instead of weekly aggregation for improved accuracy.
- Chart now displays daily data points with automatic zero-filling for missing dates to ensure complete trend visibility.

_{✏️ Tip: You can customize this high-level summary in your review settings.}

Previously WAU was calculated using non-overlapping weekly buckets. Now it uses a proper rolling 7-day window for each day, showing how WAU changes day-over-day. Definition: Users with workspace_created events on 3+ distinct days within the rolling 7-day window ending on each report date. Also updates WAUTrendChart to use 'date' field instead of 'week'. 🤖 Generated with [Claude Code](https://claude.com/claude-code) Co-Authored-By: Claude Opus 4.5 <noreply@anthropic.com>

coderabbitai · 2025-12-23T18:00:49Z

Walkthrough

This change refactors the Weekly Active Users (WAU) metric calculation from weekly aggregation to a rolling 7-day window. The backend analytics query now returns per-date data and fills missing dates with zero counts. The data contract is updated to reflect date instead of week, with corresponding frontend chart updates.

Changes

Cohort / File(s)	Summary
Frontend Chart Component `apps/admin/src/app/(dashboard)/components/WAUTrendChart/WAUTrendChart.tsx`	Updated `WAUData` interface field from `week` to `date`; aligned XAxis `dataKey` binding to match new data shape.
Backend Analytics Router `packages/trpc/src/router/analytics/analytics.ts`	Replaced weekly WAU aggregation with rolling 7-day calculation. SQL query now returns per-date WAU data. Introduced post-processing logic to fill missing dates with zero counts. Changed return payload from `{ week: string; count: number }[]` to `{ date: string; count: number }[]`.

Estimated code review effort

🎯 3 (Moderate) | ⏱️ ~20 minutes

Pre-merge checks and finishing touches

❌ Failed checks (1 warning)

Check name	Status	Explanation	Resolution
Docstring Coverage	⚠️ Warning	Docstring coverage is 0.00% which is insufficient. The required threshold is 80.00%.	You can run `@coderabbitai generate docstrings` to improve docstring coverage.

✅ Passed checks (2 passed)

Check name	Status	Explanation
Title check	✅ Passed	The PR title clearly and concisely describes the main change: switching WAU calculation from weekly buckets to a rolling 7-day window, which aligns with the primary objective of the changeset.
Description check	✅ Passed	The description provides a clear summary of changes, explains the before/after distinction, and defines WAU; however, it lacks explicit reference to the required template sections like 'Related Issues' and 'Type of Change'.

✨ Finishing touches

📝 Generate docstrings

🧪 Generate unit tests (beta)

Create PR with unit tests
Post copyable unit tests in a comment
Commit unit tests in branch fix/wau-rolling-window

Thanks for using CodeRabbit! It's free for OSS, and your support helps us grow. If you like it, consider giving us a shout-out.

❤️ Share

_{Comment @coderabbitai help to get the list of available commands and usage tips.}

coderabbitai

Actionable comments posted: 0

🧹 Nitpick comments (2)

packages/trpc/src/router/analytics/analytics.ts (1)
146-155: Potential timezone mismatch between SQL and JavaScript date handling.

The SQL query uses toDate(now()) which operates in the ClickHouse server's timezone, while JavaScript's toISOString().split("T")[0] always produces UTC dates. If the server timezone differs from UTC, the date strings may not align, causing map lookups to return 0 for valid dates.

Consider using formatDateTime(toDate(now()), '%Y-%m-%d') in SQL to match the format, or use a consistent timezone approach:
Option 1: Use formatDateTime in SQL for explicit date format
 			const { results } = await executeHogQLQuery<[string, number][]>(`
 				SELECT
-					report_date as date,
+					formatDateTime(report_date, '%Y-%m-%d') as date,
 					count(DISTINCT person_id) as wau
Option 2: Alternatively, align JS to local timezone
-				const dateStr = date.toISOString().split("T")[0] as string;
+				const dateStr = date.toLocaleDateString("sv-SE"); // ISO format in local tz
apps/admin/src/app/(dashboard)/components/WAUTrendChart/WAUTrendChart.tsx (1)
52-55: Consider updating description text to reflect rolling window semantics.

The description says "this week" but the metric now uses a rolling 7-day window ending on the most recent date. This could be slightly confusing to users expecting calendar week alignment.
Suggested text update
 					<div>
 						<CardTitle>Weekly Active Users</CardTitle>
 						<CardDescription>
-							{currentWAU} users active 3+ days this week
+							{currentWAU} users active 3+ days (rolling 7-day window)
 						</CardDescription>
 					</div>

📜 Review details

Configuration used: defaults

Review profile: CHILL

Plan: Pro

📥 Commits

Reviewing files that changed from the base of the PR and between 44723c9 and 70c1ef0.

📒 Files selected for processing (2)

apps/admin/src/app/(dashboard)/components/WAUTrendChart/WAUTrendChart.tsx
packages/trpc/src/router/analytics/analytics.ts

🧰 Additional context used

📓 Path-based instructions (6)

**/*.{ts,tsx}