1.5 KiB
1.5 KiB
graph TD
A[Add bid_history table] --> B[Add watch_count + estimates]
B --> C[Create market_indices]
C --> D[Add condition + year fields]
D --> E[Build comparable matching]
E --> F[Enrich with auction house data]
F --> G[Add AI image analysis]
| Current Practice | New Requirement | Why |
|---|---|---|
| Scrape once per hour | Scrape every bid update | Capture velocity & timing |
| Save only current bid | Save full bid history | Detect patterns & sniping |
| Ignore watchers | Track watch_count | Predict competition |
| Skip auction metadata | Capture house estimates | Anchor valuations |
| No historical data | Store sold prices | Train prediction models |
| Basic text scraping | Parse condition/serial/year | Enable comparables |
Week 1-2: Foundation
Implement bid_history scraping (most critical)
Add watch_count, starting_bid, estimated_min/max fields
Calculate basic bid_velocity
Week 3-4: Valuation
Extract year_manufactured, manufacturer, condition_description
Create market_indices (manually or via external API)
Build comparable lot matching logic
Week 5-6: Intelligence Layer
Add auction house performance tracking
Implement undervaluation detection algorithm
Create price alert system
Week 7-8: Automation
Integrate image analysis API
Add economic indicator tracking
Refine ML-based price predictions