Methodology & Sources

What MyNeta is

MyNeta is a public record of who holds office in India. It shows how they got there and what the record says. Each fact keeps a link to the source it came from. Each fact also keeps the date we fetched it.

MyNeta is not a rating site. It does not score people or rank them. It does not guess who is guilty. Where the record is thin, it says so. It never fills a gap with a guess.

Where data comes from

• Wikipedia — biographical material and comprehensive party sections are retrieved by a server-side ingestion process, pinned to a specific revision, parsed into structured facts, and cached so visitors do not repeatedly fetch the source. Content is credited "via Wikipedia" and linked to the exact source revision when available.
• Wikidata — facts such as party history, posts held and seats come from Wikidata. It is openly licensed, and we link back to the item we used.
• Roster identifiers — the current Lok Sabha and Rajya Sabha rosters and the 134-party directory are versioned seeds. Party profiles cover recognised national parties, recognised state parties, and notable registered unrecognised political parties; source enrichment is layered onto these stable identities.
• myneta.info (ADR) — self-declared assets, debts, schooling, court cases and family money come from myneta.info, run by the Association for Democratic Reforms. It draws on the Form 26 each candidate files. We fetch it when you open a profile, never in bulk. We show it only when we are sure of the match. The party and the seat must both agree, not just the name.
• PRS Legislative Research (MP Track) — how active a Lok Sabha member is comes from the MP Track dataset by PRS Legislative Research. It covers debates, questions, private bills and attendance, with national averages. The data is CC BY 4.0 and credited via PRS Legislative Research. Rajya Sabha is not in it.
• Current seat counts — the seat counts on a party page are worked out from our own list of current members. We do not fetch them from Wikidata or Wikipedia.
• 2024 Lok Sabha election results — we represent all 543 constituency outcomes. The 542 polled-seat rows carry the leading pair; most come from state and union-territory Wikipedia result tables, while an official ECI record supplies Buldhana. ECI also supplies uncontested Surat. Surat has no poll or runner-up, so its vote and margin fields do not apply. This covers the 2024 Lok Sabha election only.
• State crime statistics (NCRB via data.gov.in) — the state crime figures on the home page come from data.gov.in, India's open government data site. The data covers cases filed and the crime rate per lakh people, for 2016 to 2020. The National Crime Records Bureau put it together and it was laid before the Rajya Sabha. We show the newest year we have for a state, and we always print the year. These are past figures, not live ones. Two things matter each time you read them. First, they count crimes told to the police and written down, not crimes done. Second, how often people report a crime differs a lot by state. Better policing often pushes the count up rather than down. So we show the numbers with no ranking and no colours. They do not say how safe a state is. Where we cannot be sure which state a figure belongs to, we show nothing rather than guess.

Source classes

Every source sits in a class. The class says how close the source is to the fact. It does not say how good the source is. A court order is the record of what a court did. A news report is an account of it. Both can be right. They are not the same kind of proof.

• A1 — Court judgment / final order
• A2 — Official statutory record
• A3 — Election result / Gazette
• A4 — FIR / chargesheet / filing
• A5 — Affidavit
• A6 — Audit / CAG / committee report
• B1 — Government administrative dataset
• B2 — Parliament / Assembly record
• B3 — Company filing
• C1 — High-quality independent journalism
• C2 — International journalism
• C3 — Regional journalism
• C4 — Investigative reporting
• D1 — Academic research
• D2 — NGO / civic dataset
• D3 — Fact-checker
• E1 — Wikipedia / Wikidata / WikiBio
• F1 — Party claim
• F2 — Politician claim
• F3 — Opposition claim
• F4 — Social-media claim

Groups A and B are the record itself. Groups C to F are accounts about that record. The data quality page shows how many facts rest on each group.

Coverage in progress

MyNeta never turns a missing source into a zero or a favourable claim. The following domains are being added through source-specific ingestion and review workflows; until a record is sourced and matched confidently, the interface marks it unavailable:

• LokDhaba / TCPD (Ashoka University) — historical election results and political career trajectories require an authorised, reproducible bulk dataset rather than fragile visitor-time scraping.
• sansad.in (Lok Sabha / Rajya Sabha proceedings) — questions, debates, bills, attendance, and committee activity must be tied to official records and stable locators before publication.
• ICIJ Offshore Leaks Database — name matches are not treated as identity matches. Any future cross-reference must remain a review candidate until corroborating identifiers establish the same person.
• data.gov.in / MCA company & directorship data — company and directorship records require identity resolution, date-aware roles, and human review; a similar name alone is not evidence of a connection.
• ECI affidavit archive (affidavitarchive.nic.in) — primary Form 26 documents are preferred when a stable official document and candidate match are available; ADR/MyNeta-derived fields retain their own attribution and self-declared label.

Self-declared vs. independently verified

When we show assets or court cases, they come from the person's own sworn election form (Form 26). They are self-declared, not checked by us or by the Election Commission. A case here means it was declared. It is not proof of guilt. And if none shows, that does not mean a clean record. The form may not be sourced yet.

Allegations and accusations

An accusation is not proof. A police report, a charge, or a claim in the news means someone said a thing. It does not mean the person did it.

Each item names who made the claim, when, and where it stands now. A case that ends in a clean chit keeps its earlier steps on the record. A later step never wipes out an earlier one.

Publication boundaries

• ITR (income tax returns) — not public; legally confidential taxpayer information.
• FIR / police records — may be shown only when connected to a stable public record, clearly labelled by procedural stage, and kept separate from convictions. An allegation or filing is never presented as a finding of guilt.
• PILs, court judgments, and news-reported accusations — may be indexed when a primary record or attributable report is available. Each item must name its source, date, court or authority, procedural status, and identity-match confidence; reported accusations remain explicitly attributed and never become platform-authored facts.
• Panchayat-level data — varies widely by state and source format. It is a required coverage domain, but MyNeta publishes it only where the office, geography, term, and source can be represented reliably.

Freshness

There is no single global "last updated" date for the whole app — each fetched field carries its own fetchedAt timestamp and, where possible, a source-revision identifier. Immutable source objects are stored once, parsed facts can be reviewed independently, and scheduled ingestion checks for new revisions without making every visitor download the upstream page again.

Domain methods and safeguards