Major Hindu festivals based on Vedic Panchang (Tithi-based dates)
Dates computed for New Delhi. A festival is fixed by the tithi current at a particular moment of the local day, so the civil date can differ by location - change the city above if you are observing elsewhere.
A Hindu festival is dated by the tithi in force at the particular part of the day its rite is prescribed for, not by whichever tithi happens to be running at sunrise.
Kane states the rule in History of Dharmasastra V.i p.72: "That tithi which covers up the time (such as morning), noon, whole day &c. at which a religious act is prescribed to be done is the proper one for that act." This page applies that rule to twenty-one festival rules for any year you select, at the city you select.
In arithmetic that is an intersection of two intervals rather than a lookup. The tithi is an interval with a computed start and end. The prescribed window - purvahna, madhyahna, aparahna, pradosha, nisitha, or the day as a whole - is also an interval, a division of the local daylight or night arc. The test is whether the tithi covers the window.
Where no day is covered end to end, the fallback is the day on which the tithi merely exists in the window, then the tithi's own day, and finally, for a kshaya tithi that holds at no sunrise at all, the Hindu day it lives in. The rung actually used is named on the row, so a fallback is never mistaken for the primary rule.
The prescribed windows are proportional divisions of the local daylight and night arcs, so they open and close at different clock times in different places, while a tithi boundary is a single instant everywhere. A tithi can therefore cover one city's pradosha and miss another's on the same evening, and the civil date moves with it. That is why the page states the city it computed for instead of implying one national date.
One window on this page is deliberately not proportional. The arunodaya that governs the Naraka Chaturdashi oil bath is four fixed ghatikas, exactly 96 clock minutes, before sunrise, because that is what the almanacs count; making it proportional would let the bath drift with the season. Mixing the two units is the commonest way a festival date goes wrong by a day.
Some festivals are not one day. Divali is emitted as a sequence of separately dated rows - Naraka Chaturdashi, Lakshmi Puja on the Ashvina amavasya, Bali Pratipada and Govardhan Puja, and Bhai Dooj - because Kane's own summary is that "Taken all in all Divali spreads over about five days". Compressing that into a single date would throw away the structure the sources give it.
Where the texts genuinely disagree, the alternatives are all rendered rather than one being picked. Navaratra is shown against the seven schedules the Tithitattva lists, from the nine days of Shukla Pratipada to Navami down to Mahashtami or Mahanavami alone, plus the Devipurana's Shashthi to Dashami sequence of Durgapuja items.
Janmashtami, by contrast, is emitted as a single date, because Kane's five-rung ladder for it resolves every tie to the later day and his Janmashtami chapter never uses the Smarta and Vaishnava split he applies to Ekadashi.
The list runs from Makar Sankranti through Vasant Panchami, Maha Shivaratri, Holi and Holika Dahan, Ugadi and Gudi Padwa, Ram Navami, Hanuman Jayanti, Akshaya Tritiya, Guru Purnima, Nag Panchami, Raksha Bandhan, Janmashtami, Ganesh Chaturthi, Anant Chaturdashi, Navaratra and Durgapuja, Dussehra and Vijayadashami, Karva Chauth, Dhanteras, Divali, Bhai Dooj and Dev Uthani Ekadashi. Several of those expand into more than one dated row.
Results are grouped by month with the weekday beside each date, today is badged, and past dates in the selected year are dimmed rather than hidden, so a year reads backwards as easily as forwards. Each entry carries its type and a short description drawn from the rule that dated it, and the year selector covers ten years starting two years before the present one.
It dates observances. It does not say what a festival will bring anyone, and it does not rank one day as better than another, because the works it follows are nirnaya texts: their subject is which day an act belongs on, and nothing beyond that.
Where a classical tie-break exists but the pages settling it were not opened, both candidate days are rendered instead of one being guessed. Kane names the loci for the tie-break of a 14th mixed with the 13th and the amavasya and then writes that it "is passed over here", so both days appear and the note on the row says why.
Months are named in the reckoning each source uses, which is why Naraka Chaturdashi is the 14th of the dark half of Ashvina, or of Karttika in the purnimanta reckoning.