Ideas on how to smooth out a noisy data set

[donalmcnamee]

I've got a data set that's a little noisy and I'm looking for ideas as to how it can be smoothed-out.
 
The chart below plots the volume of a liquid in a tank on a moving vehicle.
The volume of liquid in the tank reduces as we move through the day.
 

The sensor in the tank is recording the volume approx. every 6 minutes or so.
However, as the vehicle is moving, the liquid is inclined to slosh around a bit. And if a sample is triggered when the vehicle is moving, it's inclined to read either a slightly elevated or slighted reduced reading.

I'm wondering if there's an elegant way to smooth-out these peaks and troughs in the data set?

I'm attaching an example PIBX file with an associated dataset.

Ideas appreciated.

[d_gosbell · Accepted solution]

One way to smooth out a line like this is to use a moving average.  The following code creates a moving 20 minute average. You can change the number of minutes for the moving average by changing the variable called _minsToAvg

Tank Level Smoothed = 
var _minsToAvg = 20
var _dayFraction = _minsToAvg / (24*60)
var _currentDateTime = MAX( data[EventDate])
var _startOfAvg = _currentDateTime - _dayFraction
var result = AVERAGEX (
    FILTER (
        ALL ( data ),
        data[EventDate]  > _startOfAvg 
            && data[EventDate]  <= _currentDateTime  
    ),
    CALCULATE(SUM(data[TankLevel]))
)
return  result

 The above measure produces the following output

[Rogiervanweert]

I am trying this with the data set and i get a circular dependency error?

Could you explain what i am doing wrong?

[tex628]

@Rogiervanweert  Use a measure instead of a calculated column.

/ J

[donalmcnamee]

@d_gosbell 

Thanks a million for that. Exactly what I was looking for!

Much appreciated.