My current project involves building forecasting model for asset - desktop, laptop and monitors. The data is for the last 3 years. There is good amount of seasonality in the data. There are few data points which are categorical in nature and are not consistent. Request help on how below points should be considered in the model. I am working on python environment and currently using FB's Prophet package:

1) These assets are assign to various employee which have different policies on when they can replace their asset. Would it be ok to build separate models as these incorporate different policies which allow them to change their asset hence create demand

2) The variables include assets to be forecasted at region and then country level. For example: NAMER will have US and Canada. Currently i am building separate models for each country as these are different spikes in the data demand volume

3) Most important: every year there are new product launches. This is a new demand in the data. Similarly there are assets in lets say 2017 which are no longer in production but had demand in 2017. Post Dec 2017 they were replaced with new version upgrades. When we add these variables like operating system apple, lenovo, dell etc. in the model with these conditions the model will only show demand for say the next 6 months but since there is no demand in 2018 it will show blank value for 2019. This results in data loss. What are the suggestions to deal with these type of data point. These contribute to approx. 20% of the data

4) I am currently using Prophet model in notebook, would there be some other package or methodology that i can consider knowing there are multiple categorical variables

Thank you.


Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Browse other questions tagged or ask your own question.