Log the ordinary days too. This is the part people skip and it is the part that matters. If you only open the app when your head hurts, every factor will look guilty — because you have no record of the days you had that factor and felt fine. The app compares the two, so it needs both.
Headaches are irregular, and a run of bad luck can look exactly like a pattern for a few weeks. TriggerTrace refuses to show a percentage until it has at least five days on each side of a comparison, and even then treat early results as a hint to watch, not a verdict. Two or three months of honest logging is where this starts earning its keep.
Many people find their trigger acted the day before the attack — a poor night's sleep, a skipped meal, a late glass of wine. When you look at the history, glance at the day before each attack as well as the day itself.
TriggerTrace counts days and compares proportions. It does not diagnose, it does not know your medical history, and it cannot tell you what to avoid. Its real job is to hand your doctor something better than "I think it might be cheese" — a written record of what actually happened. Sudden, severe or unusual headaches need a doctor, not an app.