产品小白尝试回答下这个产品题 可能不对 请指正。。 ..
Clarify: what time range belongs to breakfast order, is it a sudden change or it is always like this? How much less than lunch and dinner orders?
Deep dive, metric breakdown to see maybe which city/area orders are less and maybe only focus on those.. 1point 3 acres
.google и
Solution: We can build a model to predict breakfast orders, and use feature importance to have some action items
. 1point3acres
Model granularity: for example zip/city level .google и
Target variable: number of breakfast orders .google и
Features: number of available restaurants/drivers, demographics, morning traffic, delivery time, number of coffee shops(for example, if there are many starbucks nearby we can just buy those, just my guess though). 1point 3 acres
|