Dynamic Outlier Detection and Visualization with DAX in Power BI

No video

Dynamic Outlier Detection and Visualization with DAX in Power BI

Рет қаралды 12,480

Күн бұрын

Learn the theory and best methods to create a dynamic sensitive outlier detection and customize it to your particular dataset. Please find the data set and dashboard below:
github.com/Gae...
Anomaly detection can be set up by using a number of statistical techniques to set the thresholds that signal were an outlier lives. These thresholds can be set with DAX functions such as PERCENTILE.INC or PERCENTILE.EXC. There are slight differences to how these out used. However, using these, you can mimic the quartiles that exist in a traditional outlier formula.
#Outliers
#Quartiles
#DAX
#AnomalyDection
#Power BI
Contact me on Linkedin:
/ gaelimholland

Пікірлер: 19

@gregoryoliveira8358 Жыл бұрын

Thanks! I correctly calculated the limits, however, when I needed to identify the rows I had some trouble. But after I saw your video, I completely understand. Thanks again!

@absentdata Жыл бұрын

Great Job! I am glad itt helped you out.

@elrevesyelderecho 4 жыл бұрын

Excellent video. Thanks!

@praveensharma-uy7qz Жыл бұрын

Excellent...resolved my problem

@mahendraprakash2594 11 ай бұрын

Good explanation, esp using the Box & Whisker Plot. May I know how you can replicate the same analysis if there are multiple sets of data inside the table. For eg: In this case you have done for the count of users on a particular site, how to find outlier if there are 10 sites and each sites has its own set of users?

@absentdata 11 ай бұрын

If the ten sites are in a single table, you can use a column that has a indicator of each site. ie, site 1, site 2, site 3. Then you would filter by that site and the DAX formulas for outliers should adjust to that sites data giving you the specification mean, median, outliers for that site.

@mahendraprakash2594 11 ай бұрын

@@absentdata Grt, thanks for that info

@user-gh9dc2fs9y 11 ай бұрын

I did the q3,q1,iqr, max and min formulas correctly but the outlier formula shows me the following message ""A unique value cannot be determined for the 'demanda' column in the 'demanda' table." Thank you for helping me

@absentdata 11 ай бұрын

you can post your DAX here and I can evaluate it.

@miki2170 Жыл бұрын

for me, it's not working, unfortunately. I get 'outlier' in the 'Outlier formula' calculation for every value. Do you know why?

@AmanGupta-np1oc Жыл бұрын

Thank you so much

@kevinburkhardt9134 3 жыл бұрын

Could this logic be used to identify outliers within a given month, for example add a Month Index Column the far left, could we visualize that Feb has xx outliers within the Feb time frame and March has xx outliers within the March time frame?

@absentdata 3 жыл бұрын

Yes, that would work in this case since the outlier detection is based on the filtered data so you can filtered by month.

@Ide_AL 2 жыл бұрын

@@absentdata I had a similar issue. I am confident that using ALLSELECTED(outlierdata) instead on ALL(outlierdata) would make use of the slicers for months for example. Thanks a lot for the video!

@naveenkalhan95 4 жыл бұрын

my table has two columns.. date and name and then revenue amount. is it still possible? i am trying to follow your video.. but unfortunately it is not possible with two columns in the table. i there a way? thank you for your great hard work..

@absentdata 4 жыл бұрын

Yes its is possible the calculations work on individual columns so you shouldn't have an issue. I believe in your case you are using Revenue for your outlier column

@anarocioperez9483 Жыл бұрын

I am having the same problem. I have a table with date, name and sales amount. So the formula is giving me an error. Did anyone find a solution?