Analytics Platform

Build a product analytics system with these components: ## Event Ingestion API - POST /api/events - Accept batched events with validation - Fields: event_name, user_id, anonymous_id, properties (JSON), timestamp, session_id - Validate schema, reject malformed events with helpful errors - Add server-side enrichment: IP geolocation, user agent parsing, UTM extraction - Deduplicate events using event_id + user_id composite key - Queue events to Redis for async processing ## JavaScript SDK Create a lightweight (<3KB) browser SDK: - analytics.init(writeKey) - Initialize with project key - analytics.track(event, properties) - Track custom events - analytics.identify(userId, traits) - Link anonymous to known user - analytics.page() - Auto-capture page views with referrer, UTM params - Automatic session management with 30-min timeout - Batch events (max 10 or 5 seconds) before sending - Retry failed requests with exponential backoff - Queue events offline, send when connection returns ## ClickHouse Schema CREATE TABLE events ( event_id UUID, event_name LowCardinality(String), user_id String, anonymous_id String, session_id String, properties String, -- JSON timestamp DateTime64(3), received_at DateTime64(3), -- Enriched fields country LowCardinality(String), city String, device_type LowCardinality(String), browser LowCardinality(String), os LowCardinality(String), referrer String, utm_source LowCardinality(String), utm_medium LowCardinality(String), utm_campaign LowCardinality(String) ) ENGINE = MergeTree() PARTITION BY toYYYYMM(timestamp) ORDER BY (event_name, user_id, timestamp); ## Dashboard Features Build these core views: 1. **Event Explorer** - Filter/group events by any property, time range 2. **Funnel Analysis** - Define multi-step funnels, see conversion rates and drop-off 3. **Retention Cohorts** - Weekly/monthly cohorts, show return rates over time 4. **User Segmentation** - Create segments based on behavior, compare metrics 5. **Real-time View** - Live event stream, active users counter 6. **User Profiles** - Individual user timeline, all events with properties ## Query Patterns Implement these efficient ClickHouse queries: - Event counts with arbitrary GROUP BY - Funnel conversion with sequence matching - Retention matrix using cohort joins - Property breakdown with TopK - Percentile calculations for timing metrics