topic How to forecast data based on reporting month? in Reporting

How to forecast data based on reporting month?

ASISH_9 — Mon, 05 Dec 2016 09:25:50 GMT

I have a list of data named project hours and their corresponding month, like:

ProjectHours      Month
22  h             Sept
24  h             oCT
27  h             Nov etc etc

Now I need to forecast the hours for december 2016, january 2017, feb2017 up to march 2017.
How can it be accomplished?
Also, I need to subtotal the project hours under different technologies like subtotal of hours under java, subtotal of hours under dotnet. Please tell me how to do it.

Re: How to forecast data based on reporting month?

sundareshr — Mon, 05 Dec 2016 14:46:10 GMT

Have you looked at the ML Toolkit? That's probably your best bet for predictive analytics. https://splunkbase.splunk.com/app/2890/

Or, you could start with this, and tweak to get your desired results. Again, I would recommed using the app.

base search | bin span=1mon _time as month | stats sum(hours) as hours by month technology  | predict hours future_timespan=3 | appendpipe [ | stats sum(hours) by technology | eval month = "Sub-Total" ]  | sort technology