5-Day Moving Average

// Depending on the model you have...
// But let's assume that:
// You have a Dates table that's connected
// to your fact table. Say the fact table
// is Cases and stores the following columns:
// 1. Date (Dates joins on this field)
// 2. CityID
// 3. NewCaseCount
// 4. DeathCount
// There'll be other dimensions but they're
// not relevant. Of course, fact table should
// have all its columns hidden. Slicing only
// via dimensions is allowed.

[New Cases] = SUM( Cases[NewCaseCount] )

[New Case Avg (5D)] =
// change this to 7 to get a weekly
// moving avg
var __numOfDays = 5 
var __lastVisibleDay = MAX( Dates[Date] )
var __daysToAvgOver =
    DATESINPERIOD(
        Dates[Date],
        __lastVisibleDay,
        -__numOfDays,
        DAY
    )
var __enoughDays =
    COUNTROWS( __daysToAvgOver ) = __numOfDays
var __avg =
    if( __enoughDays,
        CALCULATE(
            [New Cases],
            // this will work OK if
            // Dates is marked as a
            // Date table in the model
            __daysToAvgOver
        )
    )
return
    __avg

[Deaths] = SUM( Cases[DeathCount] )

[Death Count Avg (5D)] =
// change this to 7 to get a weekly
// moving avg
var __numOfDays = 5 
var __lastVisibleDay = MAX( Dates[Date] )
var __daysToAvgOver =
    DATESINPERIOD(
        Dates[Date],
        __lastVisibleDay,
        -__numOfDays,
        DAY
    )
var __enoughDays =
    COUNTROWS( __daysToAvgOver ) = __numOfDays
var __avg =
    if( __enoughDays,
        CALCULATE(
            [Deaths],
            // this will work OK if
            // Dates is marked as a
            // Date table in the model
            __daysToAvgOver
        )
    )
return
    __avg    
    
// By the way, in M you can do anything,
// even more than in DAX, so it's certain
// that the calculation you're talking about
// CAN BE done in M.