Need help with Hierarchical forecasting

[Mozammel]

Summary:

-Every month based on the need of  our different department we ask different vendors to provide us contractors to meet our work load.

-The contractors that the vendors provide start entering FTE hours as soon as they start working.

- Based on the data we will need  a model that will predict the future FTE hours for a particular state,  a particular city, a particular job family, a particular position title.

Example:

• We have monthly data from 2014 to 2016 June for the total FTE hours entered
• The column headers are as follows:

Year	Month	State	City	Job Family	Job Title	Total Hrs
2014	April	AL	Selma	PM	PM3	10000
2015	Feb	AR	Hot Springs	Analyst	Analyst2	2300
et.c

So based on the above data we will be able to predict for

2016 September

2016 October

2016 November

For each state and for each of the city and for each of the job family and for each of the job title

Approximately how much hrs are needed.

[NeilR]

Check out the TS Factory tools in the Predictive District, and the sample here. Since you have more than one grouping field (Job Family, Job Title, etc) you may need to concatenate them into a single field using the Formula tool before feeding the data in the the TS tool.

[Mozammel]

@NeilR ,

                Thanks for the reply. I will try that. First would you tell me how would I go about concatenate in Alteryx. I do not see a concatenate command. When I trye to create a new field and adding those fields in the expressions I get this error, Parse error at Char(0): Empty. Concate resulted in a string but the field is int..

Thanks

[NeilR]

Make sure the field you're creating is a string type (eg V_WString). Then use the + operator to concatenate. Make sure each field you're concatenating is also a string type. So for example...

string_field_1 + ToString(numeric_field) + string_field_2

[Mozammel]

Hello NeiR , Thanks for the solution. It is working partially now.. However the forecast result the two columns Period and sub period are coming out to be totally incorrect. Attached is a screenshot of what I mean. Do you have any idea why this may be happening? Thanks

[NeilR]

I'm not seeing a screenshot.

[Mozammel]

NeilR

Forgot the attachment.

[NeilR]

Thanks for the screenshot but I'm not understanding the problem. Perhaps you can provide a little more detail about what you are seeing vs what you are expecting and if possible an Alteryx package that includes the workflow and some of the data.

[Mozammel]

NeilR

 Attached is the sampled data and Alteryx process flow I created. I am trying to forecast data for 2016 Jan-June based on 2014-2015 data. However if you look at the output the year is comming out to be 2017 2018 at different places.I am in need to fix that.

[NeilR · Accepted solution]

It looks like you've set your grouping field to "month" in the TS Model Factory tool. With your data, I think you want the grouping field to be "Grouped".