Finding abnormal % change from absolute figures

[Usamah22]

Hello,

i have data as below and need to calculate for each product each week what the weekly percentage change is and whether that is normal. I am struggling to calculate a % change for each week. Once i have those i hope to calculate the standard deviation and flag anything that is outside 2 standard deviations as abnormal.

Product codes	Date	Volume
a	13/01/2020	3
a	20/01/2020	4
a	27/01/2020	2
a	03/02/2020	4
a	10/02/2020	3
a	17/02/2020	2
a	24/02/2020	4
a	02/03/2020	5
a	09/03/2020	6
a	16/03/2020	3
b	13/01/2020	2
b	20/01/2020	9
b	27/01/2020	8
b	03/02/2020	7
b	10/02/2020	8
b	17/02/2020	7
b	24/02/2020	8
b	02/03/2020	6
b	09/03/2020	6
b	16/03/2020	8

[grossal]

Hi @Usamah22,

it's pretty easy to get the weekly change in %. Here is how to do it:

Configuration: 

Output:

Do you want a "moving" standard deviation or one for a and one for b? Current workflow is attached and I'll help to figure it out when I am clear about the standard deviation.

Best 

Alex

[AbhilashR · Accepted solution]

@Usamah22 - attached is an attempt to compute %change.

Hope this helps.

For standard deviation, the logic will depend on you looking to compute standard deviation at the product level across all weeks, or at a weekly level across all products.

[Usamah22 · Accepted solution]

---

@grossal wrote:

Hi @Usamah22,

it's pretty easy to get the weekly change in %. Here is how to do it:

Configuration: 

Output:

Do you want a "moving" standard deviation or one for a and one for b? Current workflow is attached and I'll help to figure it out when I am clear about the standard deviation.

Best 

Alex

---

SD for each product across weeks please. Thanks for this.

[Usamah22]

product level across weeks please

[grossal]

There we go!

Result:

Workflow:

Full workflow attached. Let me know if this is what you are looking for!

Note: I added an 'absolut' function around change, that way we also get negative abnormal values.

Alex