Change tracking between dates

[Sertan_CPH]

Look at the below example. I have a basket of fruits and I update the fruits and the price everyday. One day I have 4 fruits and the next day I am missing the orrange. Prices are also different.

Now show me how you will create a visual that shows what changed between these dates? Now imagine how to show the change trackign for the whole year. Can you help?

Date	Fruit	Price
17-09-2021	Apple	10
17-09-2021	Orange	15
17-09-2021	Banana	20
17-09-2021	Pear	25
18-09-2021	Apple	15
18-09-2021	Banana	22
18-09-2021	Pear	25

[Ashish_Mathur]

Hi,

Create a Calendar Table with a relationship from the Date column of your Data Table to the Date column of your Calendar Table.  To your visual, drag the Date column from your Calendar Table and Fruit column from your Data Table.  Write these measures

Total price = sum(data[price])

Total price on previous day = calculate([total price],previousday(calendar[date]))

Growth (%) = divide(([total price] - [Total price on previous day]),[Total price on previous day])

Hope this helps.

[Greg_Deckler]

@Sertan_CPH Here's a way to generate the necessary missing rows and calculate the change in Price:

Table17a = 
    VAR __Table =
        ADDCOLUMNS(
            GENERATE(
                DISTINCT('Table17'[Date]),
                DISTINCT('Table17'[Fruit])
            ),
            "Price",LOOKUPVALUE(Table17[Price],Table17[Date],[Date],Table17[Fruit],[Fruit])+0
        )
    VAR __Table1 = 
        ADDCOLUMNS(
            __Table,
            "Change",
                VAR __PreviousDate = MAXX(FILTER(__Table,[Date]<EARLIER([Date]) && [Fruit]<=EARLIER([Fruit])),[Date])
                VAR __PreviousPrice = MAXX(FILTER(__Table,[Date]=__PreviousDate && [Fruit]=EARLIER([Fruit])),[Price])
            RETURN
                [Price] - __PreviousPrice
        )
RETURN
    __Table1

[Sertan_CPH]

Hi @Greg_Deckler  I like your approach but I am getting an error "A table of multiple values was supplied where a single value was expected."

What can be going wrong here?

[Greg_Deckler]

@Sertan_CPH Hmm, weird, I actually tested that one. See attached PBIX beneath signature. I have to say though, I like @TomMartens approach.

[Sertan_CPH]

I don't know why it is not working. I have a huge data set with a lot of dates and they are not necessarily consequitive days. Some has one week or two weeks between them.  I have also 1000 "fruits" on my table. 

I downloaded the sample you made, it looks fine but I suspect it has something to do with filters, or dates. I am not sure. Tom's answer is good but like I mentioned, I have too many items to track on a ribbon chart. 

Do you think you can make adjustments so instead of filtering to find today's date, I would like to be able to choose two dates from a filter or something so I can compare two specific dates. 

Ps. I have a dedicated Calendar table as many of you suggested. 

[TomMartens]

Hey @Sertan_CPH ,

I consider the Ribbon chart a good fit for this data visualization task:

In regards to visualizing a complete year, I recommend creating a dedicated Calendar table, and create a relationship between the Calendar table (the one-side) and your table (the many-side). Then you can use the Calendar table on the x-axis of the ribbon chart.

This article also contains how to create a Calendar table: Time patterns – DAX Patterns

Hopefully, this helps to provide some ideas on how to tackle your challenge.

Regards,

Tom

[Sertan_CPH]

@TomMartens  Hi

It is very practical and I really like this method but as you can imagine my data is very large and not only few fruits but 1000 line items. It is simply too confusing to use ribbon chart in this case

[Greg_Deckler]

@Sertan_CPH For part of the answer, you can use MTBF, which will help you show changes between days as long as items exist. See my article on Mean Time Between Failure (MTBF) which uses EARLIER: http://community.powerbi.com/t5/Community-Blog/Mean-Time-Between-Failure-MTBF-and-Power-BI/ba-p/3395....
The basic pattern is:
Column = 
  VAR __Current = [Value]
  VAR __PreviousDate = MAXX(FILTER('Table','Table'[Date] < EARLIER('Table'[Date])),[Date])

  VAR __Previous = MAXX(FILTER('Table',[Date]=__PreviousDate),[Value])
RETURN
  __Current - __Previous