The viewparquet blog

Query Parquet on AWS S3, Cloudflare R2, GCS, or MinIO without uploading files to a third-party viewer. viewparquet streams from the bucket with DuckDB-WASM — your credentials stay in the browser.

A curated list of 11 public Parquet files you can actually download — flights, NYC taxi trips, Hugging Face NLP sets, and GeoParquet samples. Every link was tested and validated with DuckDB, with sizes, row counts, schemas, and example SQL.

Large Parquet files used to crash in-browser viewers with out-of-memory errors. Here's the engineering change — zero-copy streaming over materialized tables — that lets viewparquet open multi-gigabyte files instantly, and what it means for your workflow.

Why Parquet is the default format for fine-tuning and RAG datasets — and how to spot-check embedding columns, tokenized shards, and schema issues before a training run.

A practical walkthrough of ETL → feature store → eval sets for ML teams — with schema drift checks and in-browser spot-checks at every stage.

Discover how industry leaders leverage Apache Parquet to build scalable, cost-effective data platforms, with real-world examples and best practices for modern data engineering.

Transform complex OpenStreetMap PBF data into analysis-ready GeoParquet format, making the world's largest collaborative geospatial dataset accessible to standard data tools and workflows.

Explore how to build high-performance web maps by leveraging GeoParquet for backend processing and MapLibre GL for frontend visualization, creating responsive and scalable mapping applications.

Learn how DuckDB's in-process analytical engine paired with GeoParquet creates a powerful, serverless stack for high-performance geospatial analytics without traditional database overhead.

Discover how GeoParquet revolutionizes geospatial data storage by combining Apache Parquet's columnar efficiency with standardized geospatial metadata for unprecedented performance.