[FIXED] Is there any way to groupby and calculate statistical measures all in one step?

Issue

I know I can calculate unique count, mean, median, kurtosis, and skewness individually and merge into one dataframe, but that's a lot of steps comparing to R data.table, where you can just calculate all these in one step. Is there any way I can do a groupby and calculate all these in one step in Python as well?

df <- data[,.(ItemCount= uniqueN(Item),
             Median_val = median(Value),
             Avg_val = mean(Value),
             Skew_val = skewness(Value),
             Kurt_val = kurtosis(Value)),.(Year, Category)][order(Year,Category)]

Solution

With RootTwo's referrence link provided in the comment section, I was able to solve my own question. Note that for kurtosis, we cannot use aggfunc="kurt". It will return an error" 'SeriesGroupBy' object has no attribute 'kurt'

Below is my solution:

df = (data.groupby(['Year', 'Category'], as_index=False)
      .agg(ItemCount = pd.NamedAgg(column="Item", aggfunc="nunique"), 
           mean = pd.NamedAgg(column="Value", aggfunc="mean"),
           median = pd.NamedAgg(column="Value", aggfunc="median"), 
           skew = pd.NamedAgg(column="Value", aggfunc="skew"),  
           kurt =  pd.NamedAgg(column="Value", aggfunc=pd.DataFrame.kurt))
      )

Answered By - Jiamei