WKS / SELECTED WORK
Analysis. Investigation. Application.
Questions worth
following.
Data analysis and investigative tools that turn complex information into work you can examine.
Explore the work ↓
Python / SQL / Applied analytics
Featured case study
01 / 05
POSTGRESQL / MARKETPLACE INTEGRITY
SQL Risk Lab
From marketplace data
to an investigation queue.
Reusable SQL investigations, explainable review signals, and a public Olist case study.
99,441orders examined
15investigation queries
Read the case study ↗
Olist marketplaceOrder review
8,349orders queued for review
Medium priority or higher8.4%
- Total orders
- 99,441
- Sellers
- 3,095
Priority reflects review signals, not confirmed misconduct.
More selected work
02 — 0502 / CASE STUDYPublic Bitcoin transactions become a review queue, flow map, and analyst memo, with attribution limits made explicit.
BitcoinObserved flow excerpt
Listed sample wallet15FcqYRb…pycGch
Observed next hop18fPjXb6…aS5q7A
Unlabeled onward output19kWDy5x…w3AKu8
Public-chain observations. A transfer does not establish common ownership.
03 / CASE STUDYPublic player and pitch data used to examine development, compare cohorts, and test translation across levels.
Player developmentCurated cohort
Successful MLB outcomes
Selected organizations; small samples. Outcome labels follow the case-study definitions.
04 / CASE STUDYA public-safe workflow connecting source intake, entity extraction, travel advisories, and analyst reporting.
Public-source intakeRSS case study
240source rows
→149queued items
- Normalize
- Correlate
- Review
Analyst outputsCasepack / Brief / SITREP
05 / CASE STUDYSynthetic and redacted evaluation cases organized into risk clusters, review queues, and documented escalation decisions.
AI evaluationSynthetic fixture
- Cases
- 18
- Risk clusters
- 12
- Register entries
- 7
Critical queue9 cases
Evaluation health8 flags
Synthetic, redacted cases with documented reason codes.