Disclosure: I built this and I host it. Free, CC-BY 4.0, no signup, no API key.<\/p>\n
What it is: every date in 2026 and 2027 evaluated against the classical Hindu electional-astrology ruleset, for eight life events (marriage, griha pravesh, vehicle purchase, property, business, signing, travel, general).<\/p>\n
- \n
- muhurat-2026: 1,291 dates<\/li>\n
- muhurat-2027: 1,324 dates<\/li>\n
- CSV and JSON, both years: https:\/\/bela.so\/datasets\/<\/a><\/li>\n<\/ul>\n
The part that might be useful beyond the religious use case: each record carries the per-rule breakdown, not just a score. So for any date you can see which classical screen it passed and which it failed \u2014 tithi fitness, nakshatra suitability, yoga exclusions, panchaka, prohibited months, and so on. It’s a labelled ruleset-evaluation dataset as much as a calendar.<\/p>\n
How it was computed: Swiss Ephemeris (full DE-series files), Lahiri\/Chitrapaksha ayanamsa, mean nodes, drik-ganit \u2014 true solar\/lunar positions rather than the mean-motion tables a lot of published almanacs still use. Deterministic: same input, same output, and the same engine version reproduces the files byte for byte. No language model touches any number in there. Method is written up at https:\/\/bela.so\/methodology\/<\/a> if you want to check the approach before trusting the data.<\/p>\n
Caveats, so nobody gets surprised:<\/p>\n
- \n
- Ayanamsa is a choice, not a fact. These are Lahiri. Raman or KP would shift every boundary by a few arcminutes, which moves some date boundaries.<\/li>\n
- Times are computed per-city from local sunrise. The dataset ships a reference set of cities; it is not every town on earth.<\/li>\n
- “Auspicious” here means “passes classical rule X” and nothing more. The rules are the object of study; I’m not making a claim about outcomes.<\/li>\n<\/ul>\n
Happy to answer methodology questions or add fields if there’s something obvious missing.<\/p>\n<\/div>\n
submitted by \/u\/National_Engineer277 <\/a>
[link]<\/a><\/span> [comments]<\/a><\/span><\/p>
