Anomer

Anomer tells a small shop owner why a number changed, and how much each cause mattered. Ask "Why did orders drop Oct 12–18?" and get a breakdown: ads went quiet −125 orders, then Instagram, steady growth and weather, plus what could not be explained, each with a 95% range. A small shop's numbers live in different places (orders, Meta ads, Instagram, the weather), and nobody there has an analyst to connect them. Anomer shifts every connected number 0–7 days, puts it next to the day of the week, the day of the month and the trend, and fits one Lasso regression with time-ordered cross-validation. It then splits the change between two periods cause by cause. Nothing about paydays or weekends is hard-coded. Muse Spark reads the question and explains the result; every number comes from code. The demo shop uses real NOAA weather. Its ads, Instagram and orders are simulated with planted causes, so the answer can be checked.