Home>Projects & Open Datasets

Projects & Open Data Initiatives

Orgpedia designs extraction technology, benchmarks, and interactive public portals to convert unstructured government orders into accessible, machine-readable civic knowledge. All data and tools are released under open licenses.

Ministers of India
Flagship PortalLive & Continuously Updated

Ministers of India (1947–To Date)

A systematic, interactive exploration of every minister, prime minister, deputy prime minister, and cabinet portfolio in the Government of India from Independence (15 August 1947) to the present day. Built by parsing hundreds of Cabinet Secretariat gazette notifications, extracting officer appointments, portfolio re-allocations, and building continuous tenure graphs.

History Span77+ Years
Orders Processed904 Documents
Verified Tenures3,885+ Records
Languages22 Scheduled
Maharashtra GR Summary
State GovernanceActive Weekly Monitor

Weekly Summary of Maharashtra Government Resolutions (GRs)

State governments issue thousands of Government Resolutions (GRs) every month across departments like Finance, Home, Revenue, Public Health, and School Education. Orgpedia automatically ingests these documents from the official portal, extracts subject metadata and financial sanctions, and produces weekly digestible summaries for civic analysts, journalists, and policy researchers.

Total GRs Tracked100,000+ GRs
CadenceWeekly Digest
LanguageMarathi & English
Issuing AuthorityGovt of Maharashtra
indicGovBench-Orders
Benchmark DatasetResearch Release

indicGovBench-Orders: Benchmarking AI on Indian Government Documents

Indian government orders have distinct peculiarities: degraded historical scans, unruled tables, mixed digital and handwritten text, and Devanagari/Indic typography. indicGovBench-Orders is a specialized evaluation benchmark spanning Hindi, Marathi, and English orders, designed to empirically test how commercial and open-weight Vision-Language Models (VLMs) and OCR engines perform on real-world civic administration orders.

Orders Evaluated300 (Scaling to 1,000)
Primary LanguagesHindi, Marathi, English
Key FocusUnruled Table Parsing
Expansion7 Additional Scripts