Query the Data — Data Darbar

How this works. There is no server. The data sits in Parquet files on the CDN and the query engine (DuckDB-WASM) runs inside your browser tab, so nothing you type is sent anywhere. Full DuckDB SQL is available — CTEs, window functions, corr(), PIVOT. Click a table name to start a query; click the row to see its columns and caveats. Queries are shareable: the URL carries your SQL.

Use the same data from Python, R or the DuckDB CLI

# The Parquet files are public static assets — query them over HTTP, no download step.
import duckdb
duckdb.sql("INSTALL httpfs; LOAD httpfs;")
duckdb.sql("""
  SELECT name, prov, mpi
  FROM 'https://hibasameen.github.io/datadarbar/data/warehouse/mpi_districts.parquet'
  WHERE low_n = 0 ORDER BY mpi DESC LIMIT 10
""").show()

# R
# arrow::read_parquet("https://hibasameen.github.io/datadarbar/data/warehouse/mpi_districts.parquet")

# The table list, schemas and caveats live in
# https://hibasameen.github.io/datadarbar/data/warehouse/catalog.json