JM

05 / Personal projectData product

LEGO Resale Intelligence

A tool that compares a listing with the market, works out the margin and says how far to trust the figure.

Try the public demo

Type
Independent personal project
My role
Conceived and built it, with AI
Product
FastAPI back end and Next.js interface
Status
Public demo with demo data

Personal project, unconnected to the LEGO Group. The films show the product demo with demo data. The app’s interface is in English; the briefings it generates are in Spanish.

01 / The problem

A price alone isn’t an opportunity.

Buying a retired set to resell means checking a lot by hand.

The decision isn’t just “it’s cheap”. It’s margin, risk and timing.

What to check in each listing

  1. 01Asking price
  2. 02Shipping cost
  3. 03Condition of the set
  4. 04Pieces, box and instructions
  5. 05Fair market price
  6. 06Margin after fees
  7. 07Risk of overpaying or buying incomplete

02 / The question

What would I need to understand before considering an opportunity?

03 / How it works

Four steps, from a listing to a decision.

Analyse · Signal · Track · EvidenceDemo data

04 / Data governanceReal project data

Every set, with its provenance. The one that doesn’t add up, under review.

Rather than presenting the catalogue as verified market data, I added each set’s provenance and an audit log.

A set only counts as verified if it passes all 8 checks. Tap any set to see its own.

49/50verified

Star Wars17

Modular Buildings15

Icons6

Ideas7

Technic4

Creator Expert1

VerifiedUnder review

Icons10220

Volkswagen T1 Camper Van

Under review · 3 of 8 in doubt

  • Release year
  • Retirement year
  • Official price
  • Piece count
  • Popularity
  • 2 data sources
  • 2 price sources
  • 3 price snapshots
Retirement year
2012, 2020 or 2022, depending on the source
Pieces
1,332, 1,333 or 1,334, depending on the source
Official price in €
No stable reference

Brickset, BrickRanker and BrickEconomy disagree. Forcing a verification would have undermined everything else, so it stays as demo data.

05 / My role

The questions and the judgement are mine.

I conceived and built the product with intensive AI support. I defined the problem, the business logic, the quality criteria and the evidence rules. AI helps me develop and iterate.

What I did

  • I defined what someone needs to know before buying, and in what order.
  • I designed and built the product: analysis, tracking, market and evidence.
  • I audited the 50 catalogue sets against external sources.

What it isn’t

  • A live pricing platform: the price history is demo data.
  • A LEGO Group product, nor endorsed by it.

06 / Figures

116Automated tests
50Sets in the catalogue
250Price snapshots
49/50Sets verified

Figures from version 2 of the project. The price snapshots are demo data, not live market prices.

07 / Decision value

The most important decision was not to overstate the data.

Learning

Saying what is under review makes everything else more credible.

Independent personal project. LEGO® is a trademark of the LEGO Group, which does not sponsor, authorise or endorse this project. The price data is demo data.