Aggregation and axes

Summarising an array

The aggregations are methods on the array, and each has a NumPy function twin:

a.sum()   a.mean()   a.min()   a.max()   a.std()
a.argmin()   a.argmax()      # the POSITION of the smallest / largest

For a two-dimensional array, axis decides which direction to collapse. This is the part worth slowing down for:

m = np.array([[1, 2, 3],
              [4, 5, 6]])

m.sum()            # 21 — everything
m.sum(axis=0)      # array([5, 7, 9])  — down the columns
m.sum(axis=1)      # array([6, 15])    — across the rows

Read axis=0 as "collapse the rows, leaving one value per column". People remember it as "axis 0 is vertical", which works as long as you also remember the result has one entry per column.

Your turn: sales has one row per shop and one column per quarter. Find the total per shop, the total per quarter, and which shop sold most overall.