Rohit has given the correct answer, namely that the WL is case sensitive with all inbuilt functions starting with a Capital letter and all concatenated names having each new word begin with a Capital. The problem with the data returned by FinancialData is that it is not regularly spaced and hence cannot be used to estimate the underlying process.:
In[1]:= RegularlySampledQ[FinancialData["EUR/USD",{{2021,5,1},{2021,9,30}}]]
Out[1]= False
This can be accomplished either by using the TemporalData[data]["Path",All, period] method above, or in two other different ways:
either by using
data=TimeSeriesResample[FinancialData["EUR/USD",{{2021,5,1},{2021,9,30}}]]
or
data=TemporalData[FinancialData["EUR/USD",{{2021,5,1},{2021,9,30}}],TemporalRegularity->True]
Michael