Found21
Fixed + verified21
Held for approval0
Recall1.00
One offline run on demo data with planted problems (recorded platform test, November 2026).
How RetailAnalytics heals data, step by step
INData inKaggle carrie1/ecommerce-data
01Detect21 problems found
02Diagnose4 problem types named
03Plan21 safe to fix · 0 need a person
04Fix + verify21 fixed and re-checked
05ReportPlain-English report and Control Tower record
- Client problem
- Owners never see a clean daily picture; guest checkouts and bad prices distort numbers.
- What it does
- Fixes missing customer IDs, impossible prices and duplicates; builds daily sales; detects revenue drops.
- How it heals
- Repairs data and has a CrewAI team write a grounded daily brief.
- Human gate
- None (read-only insight).
- Typical buyer
- Retail owners
- Built with
- LangGraph self-healing loop with a CrewAI crew (analyst, fixer, reviewer)
- Demo data
- Kaggle carrie1/ecommerce-data (UCI Online Retail, ~540k invoice lines).
- Tables
- online_retail ← Kaggle carrie1/ecommerce-data
- Target outcome
- Target: owner brief ready by 6 AM every day.
- Path to production
- Insight Pack.
Issues detected on the demo data
| Issue | Risk | Proposed fix | Count |
|---|---|---|---|
| missing_customer_id 'customer_id' is empty | low | Tag as guest customer | 10 |
| out_of_range_price price=-1.69 outside [0, None] | low | Quarantine – likely cancellation or entry error | 3 |
| duplicate_record Duplicate on line_id | low | Drop duplicate invoice line | 5 |
| revenue_drop Revenue 1,843 vs 7-day average 3,575 | low | Explain in the daily brief with drivers | 3 |
How we demo it
- Load UCI Online Retail II.
- Run RetailAnalytics.
- Read the daily brief.
Built on AWS
Runs in the platform container on EC2 Graviton (or on demand in Lambda), writes healed tables to the S3 lakehouse, uses Amazon Bedrock only to write explanations, and reports six metrics to CloudWatch.
Step into DigiQuench.
Agents that heal your data.
Watch an agent find planted problems, fix the safe ones and wait for your approval on the rest. Thirty minutes, live.