Question about Forecasting

[Anonymous]

Hi

I’m trying to forecast the sales for the coming 3 months. I’ve 4 years of data. The thing is that I’ve two months each year where there is spike in sales.

I’m trying to use the Forecast function in Power BI to forecast the sales in the coming 3 months. The problem is that if I used the last 15 months to forecast the coming 3 months, the spikes in two months makes the results not accurate.

What can I use in order to reduce the impact on the spike in the two months.

[grantsamborn]

Hi

I noticed your post yesterday but noticed you hadn't any replies...

So, I have a few questions:

What is the granularity of your data?
If Daily, I would try to use a Rolling Average measure.

BUT your chart seems to show monthly but a monthly axis (Yr-Mon) would be "categorical", making forecasting impossible (as far as I know).

Any comments?

(*Fake data* would probably help.)

Edited to add:  Consider setting Seasonality at 365 data-points.