I have sales data of several refurbished smart phone devices. The sales occurs daily but not all phones are sold daily.. How do i calculate price elasticity of this data.. the sales are wholesale sales and there doesnt appear to be a linear relationship between the price & quantity at aggregate level. SHould I separate out individual models? SHould i try some time series analysis on this data. multi linear regression is giving very bad outputs.. and how do i include depreciation?