High-level explanation:
Surge (dynamic) pricing raises fares in zones/times where rider demand outstrips available driver supply to incentivize drivers and balance the marketplace. The system ingests real-time signals (incoming ride requests, available drivers, estimated ETAs, historical demand patterns) and calculates a multiplier per geo-tile using supply-demand ratio, wait-time targets, and business rules (caps, fairness constraints). Multipliers update frequently and are published to riders/drivers.
Three BI metrics/visualizations to monitor surge fairness and rider impact
- Geographic Surge Heatmap with Equity Overlay
- What: Choropleth of current and historical surge multiplier by tile with overlays for rider sociodemographic segments (e.g., income bands) and average wait time.
- Data sources: Real-time pricing feed, trip requests, driver availability, census/zipcode socio-economic data, ETA logs.
- Alert rule: Trigger if any tile shows multiplier >1.8 for >30 minutes AND average wait time for low-income zipcodes in tile > median city wait time by >50%.
- Rider Cancellation & Complaint Rate vs. Multiplier Scatter
- What: Scatter/time-series showing rider cancellation rate and “price-related” complaint volume against surge multiplier bins; include cohort filters (new vs. repeat riders).
- Data sources: Trip lifecycle events (requested → accepted → cancelled), support tickets tagged for pricing, user profile.
- Alert rule: Trigger if cancellation rate for multiplier bin [1.5–2.0] increases by >2ppt week-over-week AND complaints per 1k trips in that bin >5.
- Fare Elasticity & Access Loss Metric (Rides Lost Estimate)
- What: Time-series estimating rides lost due to surge using baseline demand model (predicted requests at no-surge) vs. observed completed rides, with elasticity estimates by segment.
- Data sources: Historical demand model, A/B experiment logs, completed trips, search/request logs.
- Alert rule: Trigger if estimated rides-lost >3% of hourly demand in a region AND elasticity implies >10% drop in trips per 0.1 multiplier increase for vulnerable cohorts.
Each visualization should support drill-down by time, region, rider cohort, and include causal context (driver shortages, events). Alerts route to pricing ops and policy/UX leads with automated snapshot and suggested mitigations (cap multiplier, driver incentives, communication).
