Pandashells lets you use Pandas from the bash command line. It allows you to combine unix command-line tools (awk, grep, sed, etc.) with the power of Pandas Dataframes and Matplotlib visualization.
Here is a one-liner that performs the exact same aggregation demonstrated by the example consecution pipeline.
cat sample_data.csv | \
p.df 'df["group"] = ["adult" if a>=18 else "child" for a in df.age]' | \
p.df 'df.pivot_table(index="group", columns="gender", values="spent", margins=True, aggfunc=sum).fillna(0)' \
-o table index