I have historical sales 2 years of data for multiple customers and products at daily level and objective is to calculate price elasticity of demand for these customers. Variables in data are - Customer, Product, Sales Date, Demand, Unit Price. I have used traditional lol-log linear model to calculate the price elasticity of demand. I have run regression model for each customer-product combination separately in order to calculate the price elasticity at customer-product level. End Objective is to use this price elasticity for profit maximization exercise

Model Equation : Log (Demand) = b1 * log (Price)+e

I have following questions:

- Does my methodology make sense? If so, then price elasticity of demand would coefficient b1?
- I have read that, price elasticity should be negative. What if coefficient b1 is positive?
- With this data, how do I consider endogeneity of variable?
- How can I perform similar exercise mentioned on this website ?

I would appreciate any kind of help to solve this problem Thank you!