We have data on product sales at a monthly level of aggregation, with up to 5 years worth of data (varies by product) in a single csv file. Some forecasts are strongly seasonal, others a linear trend and others are fairly uniform. We require a robust monthly forecast for each product for each of the next 12 months. There are about 650 products and a total of about 15500 records. Note data has missing records where we sold zero, or where products were introduced or ceased part way through (treat these as zero). We want a batch program written in something like R or python that will produce sensible forecasts without intervention. There will be little human intervention. We would like a Holt Winters model with some statistical testing of parameters so that we adopt seasonality, trend if we find it, not otherwise. I believe there are libraries in R that can do this. I can look at re-formatting data if required. The output format is more fixed. We would also like a measure of forecast error (standard deviation) along with central forecast. I will post the actual data file to pre-qualified freelancers. We would also like separate program that can output a file with weekly values from the monthly file (a simple subdivision based on the number of mondays in each month - equal weekly value to be the same within a given month.
35 freelancers are bidding on average $464 for this job
I can implement holt winters using R , however I'd prefer to work with a small subset. Also I can create an application in Shiny to display predicted results
Hello! I am a python developer. I looked at your project and it seems interesting. I have all necessary skills required for this project. Ping me to discuss in detail.
Hello I am interested in your project. I was working for Machine learning engineer for 2 big electrical company in Germnay. Projects were also forecasting electrical and gaz consumption. Best regards Houssem