Daily Forecasting

Hi all, could someone please help me with a daily forecasting tool. I have attached some dummy data for orders despatched for couple of countries by products, by day for the last few months. Based on this data I would like to be able to forecast orders for future dates, by country, by product. For example, I would like to forecast orders from USA, for Sunday 1st Oct 2017 based orders of Sunday 2nd Oct 2016, by taking into account any increase or decrease on Sundays over the last 12 months, but excluding period of Nov-Dec, listed as a winter increase. There is data for many countries to select from, so I need to be able to select any country and forecast for the next few days, specific for the country. I would really appreciate if someone could please help. thanks

Hi gsrai31,

This may not be what you want, but it's the closest I can get.
There may be someone else that will come along with a better solution.
Good Luck with your project!
Thanks

from quickly looking at the numbers it seems like you don't have the information here required to forecast very accurately.
There is probably an unrecorded factor that you may well know driving it like weather or staff hours, advertising spend or something.
Or possibly you just need several more years of data.

Gsrai31,

No problem, thanks for the reply.
Good Luck with your project and I as you do, look forward to any replies that help answer your question!

Try experimenting with FORECAST.ETS() to start out (ETS portion stands for Exponential Triple Smoothing).

For any type of forecasting, you will need to test out various models and validate. Then there's amount of accuracy required and how much time/cost you are willing to spend on the process of developing and maintaining the model.

Harvard Business Review has many good articles on it (I think you get 3 free articles).

Here's link to Open-Access Textbook on Forecasting.
https://www.otexts.org/fpp

You may find while Excel gives great starting point, it may become bit cumbersome to maintain. R is very powerful language for statistical computing.
Here's sample links to how R can be used in forecasting.

http://www.datamic.net/blog/dynamic-...power-bi-and-r
https://www.datascience.com/blog/int...ence-tutorials

EDIT: Fore seasonality forecast, you will need minimum of 2 years data. Preferably more. Without seasonality, you can get away with less.

That function is new in Excel 2016 and not available in 2013: https://support.office.com/en-gb/art...6-625ccaaa60b4

You can do few different things.

1. Install data mining add-in from MS. It has forecasting tool using ETS ARIMA algorithm.
2. Integrate R script in Excel and use forecast analysis package
3. Do forecast in R studio and markdown

Other than that... you can do analysis in PowerBI using R. If you search for "Excel 2013 ETS forecast" you may find some other methods.