Thank you for that. I was doing it in the research environment. Do you know how to make it work there?

The issue is your assignment statement. Nothing to do with the factors, filters, window lengths, or masks. I've added comments to the code below which steps through the problem.

def make_pipeline():  
    # The code below creates a new undefined pipeline and sets the variable 'pipe' to reference this object  
    pipe = Pipeline()  

    # The following code creates a universe filter and a couple of factors  
    # All looks good with these  
    universe=Q500US()  
    daily_returns = DailyReturns(mask=universe)  
    returns = Returns(window_length=2, mask=universe)  

    # The following creates a new pipeline with the defined columns and screen.  
    # It works as expected BUT the pipeline object isn't assigned to anything.  
    # It gets created but immediately deleted by Python because there is no way to reference it  
    Pipeline(  
        columns={  
            'daily_returns' : daily_returns,  
            'returns' : returns  
    },  
        screen = universe)  

    # The function returns the first undefined pipeline 'pipe' which isn't what's intentioned  
    return pipe  

The fix is simple. Eliminate the initial 'pipe' assignment and, instead, assign 'pipe' to the second pipeline.

def make_pipeline():  

    # The following code creates a universe filter and a couple of factors  
    # All looks good with these  
    universe=Q500US()  
    daily_returns = DailyReturns(mask=universe)  
    returns = Returns(window_length=2, mask=universe)  

    # The following creates a new pipeline with the defined columns and screen.  
    # Ensure it gets assigned to the variable 'pipe'  
    pipe = Pipeline(  
            columns={  
                'daily_returns' : daily_returns,  
                'returns' : returns  
            },  
            screen = universe)  

    # Now return 'pipe' which is correctly defined  
    return pipe  

That should do it.

Thank you so much Dan!