Notes

Long-form technical writing on statistics, data architecture and the systems that carry them into production.

1 September 202646 min read

Voluntary Tax Compliance as a Measurement System: Evidence, Identification, Prediction, and Forecasting

One compliance rate hides four different statistical objects. Separating measurement, causal inference, risk scoring and forecasting, with code that runs.

StatisticsData scienceEconomics

8 August 202616 min read

Real-Time Geolocation Backends: PostGIS Spatial Indexes, FastAPI Streaming, and Docker Dispatch

Splitting the write firehose from the latency-sensitive read: Redis GEO for the live fleet, PostGIS for durable spatial queries, and SSE for dispatch.

ArchitectureBackendDevOps

6 August 20267 min read

Production RAG Architecture: Next.js App Router, PostgreSQL HNSW Indexes, and Half-Precision Vector Streaming

Half-precision vectors to stay inside PostgreSQL page limits, SQL-native Reciprocal Rank Fusion, and streaming past reverse-proxy buffering.

ArchitectureAI & RAGBackendFrontend

6 August 20263 min read

Dockerizing R MCMC: Scalable Bayesian Inference in the Cloud

Packaging an R MCMC environment into a Rocker image, then scaling it across a Kubernetes cluster without losing reproducibility.

DevOpsStatistics

5 August 202610 min read

Architecting Distributed MCMC Pipelines: Go gRPC, pgvector, and Next.js Server-Sent Events

Concurrent sampling in Go, binary transport over gRPC, posterior state in pgvector, and convergence metrics streamed to the browser.

ArchitectureBackendStatisticsFrontend

5 August 20263 min read

A Multi-Language Data Science Architecture: Uniting Python and R in Positron IDE

Python prepares the data, R visualises it, and Quarto binds both into one reproducible report, all inside a single IDE.

Data scienceStatisticsArchitecture