Indian Credit Card Facts — an open dataset

A dated, machine-readable record of the Indian credit-card market: what each card charges, what it earns and where that earn is capped, how its terms have changed over time, and where its points can be transferred — plus one realistic point value per card. Four tables, JSON and CSV, release v2026-08-16 (schema v3): cards — 280 cards. Fees, stated eligibility, reward rates with their caps and one point value per row, perk types, network and tier. changes — 1617 events. Dated changes to card terms, each with direction, magnitude, before/after values, and at least one external source. transfers — 153 edges. The points-transfer network: source programme, destination programme, ratio, tier gate, cap, transfer time, source. valuations — 233 cards. One realistic point value per card and per programme, in paise per point, on the default redemption route. The line this dataset draws. It carries facts — and it does not carry CardAdvisor's editorial model: ratings, rupee valuations of perks and welcome bonuses, route-by-route point values, or valuations of airline and hotel currencies. Those are opinions and are published on the site where they can be read in context. Facts are open under CC BY; the model is shown, not shipped. Erratum for v2026-08-15. The first release's files carried several of those model columns (expertRating, welcomeBonusValueRs, perk values, route-level point values, transfer-graph currency values) while its datasheet said they were excluded. The datasheet stated the intent correctly and the files did not honour it. That release stands as published; this release (schema v3) is the correction, and the erratum is recorded in the release index. There is no public structured record of Indian credit-card terms. Issuer pages state today's terms and nothing else; when a reward rate is cut, the previous terms disappear from the web. This dataset exists so that questions like "what did this card pay last year, and when did that change" have somewhere to be answered. Known accuracy is published, not asserted. A fixed sample of 30 cards across 15 issuers is re-checked against issuer primary sources and the failures are published alongside the successes at cardadvisor.in/data/audit. On the 2026-08-15 pass, joining and annual fees were corroborated on every readable source (24/24 each); the monthly APR column was corroborated for only 3 of 22, because APR is published in issuers' fee schedules rather than on card pages. Anyone using the APR column should treat it as the weakest in the dataset. Contains no personal data of any kind — it describes financial products, not people. Coverage note: the cards table aims at the complete population of actively marketed consumer credit cards in India. The changes table is explicitly not complete — it is bounded by what was publicly reported, is strongest from 2023 onward, and thins the further back it goes. Any analysis of change frequency over time must account for that observation bias. Full datasheet, field dictionary, licence and errata policy: cardadvisor.in/data. Immutable releases: cardadvisor.in/data/versions. Independent: no issuer, network or bank funds, sponsors, reviews or has any input into this dataset.