Card is Not Accurately Counting Distinct Rows

[henriley]

Hello, I'm realtively new to PowerBI and am struggling to understand some nuances of the platform. This is the view of the current iteration of my dashboard:

The highlighted slicer is used to filter the cards in the top right based off various feature values. For example, if I set the slider to be 0-100 for Feature 1, the only surveys that would be counted are the surveys that had a value that lands in that range.

I create an index in each row to mark each unique survey instance. The # of Surveys card is counting the number of unique indices in a table using this measure:

# of Surveys = DISTINCTCOUNT('Fact Features'[Index])

When I set the value slider to be between -150,000 and -9000, I would expect the # of Surveys card to read 9, not 10, based off the data view set to have the same filters (<=-9000 for the feature value and the feature name set to Feature 1):

What it seems like is the card is double counting the indices if a specific Business Partner has two different surveys with a number that falls within the range and a number that falls out of the range. It will automatically count both even if both numbers don't fall in the range. The table below is in the report view with Feature 1's value set to be <=-9000:

We can see that riley is being counted twice even though he only has one value in the range.

Any help would be appreciated on how to get the accurate counts.

[PaulDBrown]

please share some sample data or PBIX file using OneDrive, Dropbox or similar, and a depiction of the expected outcome. Thanks!

[henriley]

This is the raw data that is filtered/counted in the cards:

[PaulDBrown]

Thanks, but we need data, not an image, to be able to work on!

[henriley]

Thanks PaulDBrown, my company does not allow me to post a link that allows direct access to the PBIX/Raw data. The best I can do is posting tables here.

Dim Geography

bpartner_id	premise_id	latitude_derived	longitude_derived	move_in_date	move_out_date	city_bpar	postalcode
bob	1	42.59364	-83.0748	1900-01-01T00:00:00.000+0000	9999-12-31T00:00:00.000+0000	traverse city	49684
lori	2	42.60364	-83.1448	1900-01-01T00:00:00.000+0000	9999-12-31T00:00:00.000+0000	sterling heights	48310
maggie	3	42.44364	-83.2248	1900-01-01T00:00:00.000+0000	9999-12-31T00:00:00.000+0000	troy	49459
riley	4	42.59364	-83.0748	1900-01-01T00:00:00.000+0000	9999-12-31T00:00:00.000+0000	sterling heights	48310
erin	5	42.35364	-83.0188	1900-01-01T00:00:00.000+0000	9999-12-31T00:00:00.000+0000	farmington	48334
debra	6	42.47161	-83.5062	1900-01-01T00:00:00.000+0000	9999-12-31T00:00:00.000+0000	novi	48374
sharon	7	42.70364	-83.3548	1900-01-01T00:00:00.000+0000	9999-12-31T00:00:00.000+0000	port huron	48060
pete	8	42.42364	-83.5148	1900-01-01T00:00:00.000+0000	9999-12-31T00:00:00.000+0000	birmingham	48009

Fact Features/Feature Selector (1)

nps_bpartner_id	survey_nps_raw	nps_classification	survey_date	Water_Only	Heat_Only	Combo	is_happy	is_adult	feature_1	feature_2	feature_3
riley	8	passive	2018-02-19T20:15:06.000+0000	1	0	0	FALSE	TRUE	-10600	50	100
erin	10	promoter	2018-03-03T15:48:22.000+0000	0	1	0	FALSE	TRUE	-150000	25	110
riley	8	passive	2019-04-13T18:17:38.000+0000	1	0	0	FALSE	TRUE	-8000	20	120
erin	7	passive	2019-03-24T11:28:24.000+0000	0	1	0	FALSE	TRUE	-70000	30	130
bob	10	promoter	2018-12-04T13:53:48.000+0000	1	0	0	FALSE	TRUE	-20000	40	140
lori	0	detractor	2018-08-07T10:16:02.000+0000	1	0	0	TRUE	FALSE	-30000	70	150
maggie	9	promoter	2019-07-19T12:38:18.000+0000	0	0	1	TRUE	FALSE	-40000	60	160
debra	9	promoter	2019-03-24T13:02:53.000+0000	0	1	0	TRUE	FALSE	-50000	55	170
sharon	3	detractor	2018-05-22T08:49:32.000+0000	0	0	1	TRUE	FALSE	-60000	45	180
pete	9	promoter	2018-03-20T17:04:52.000+0000	0	1	0	TRUE	FALSE	-70000	85	190

Let me know if this makes it easier to assist.

The other idea I have is the potential relationships I have between the tables:

Feature Selector (1) is connected to Dim Geography on a Many-to-One relationship through the Business Partner. The Business Partner can take multiple surveys but will only have one latitude and longitude in Dim Geography. Dim Geography connects to Fact Features on a One-to-Many relationship on the Business Partner. The one location connects to the many surveys taken on the Business Partner. Fact Features and Feature Selector (1) are duplicates of each other. Feature Selector (1) is used to filter the values of the different features and Fact Features is used for the counts/percents we see in the cards on the top of the report.