Project Introduction - Simzworld/Business_Forecasting_For_Car_Sales GitHub Wiki
What are we forecasting and the term?
We are forecasting unit volume demands trends for consumer automobile vehicles in the United States based on historical sales data. We think this is an interesting data set to explore especially in today's environment considering the supply constraints tied to the global pandemic, a high influx of liquidity due to government cash injections, and a mass shortage inventory. We would like to forecast a "short-term" outlook for the next 6 months
How are we going to measure the accuracy of the forecasting model?
a. Splitting data into test and training sets based on an 80 – 20 % split. In our case, this would be splitting the sales data based on the number of years
b. Comparing the error measures of the model with other basic models like Naïve and seasonal Naïve
c. Measuring residual percentage errors by calculating the Mean absolute percentage error: MAPE
d. There is a possible fourth option to running the model on car sales for a different brand comparable to the one we are using which in our case is Honda
Source Data Details
The data set consists of Honda unit volume sales. The data is displayed on a monthly basis from 2005 through August 2021. The data is sourced from "goodcarbadcar.com" where they have a tabular format of the whole sales data.
Links to group members profiles
https://github.com/harshalutekar/Honda-Sales-Forecasting/projects