8.1. Statistics Reduction
8.1.1. SetUp
>>> import numpy as np
8.1.2. Sum
a.sum()np.sum()np.nansum()
>>> a = np.array([[1.1, 2.2, 3.3],
... [4.4, 5.5, 6.6]])
>>>
>>>
>>> a.sum()
np.float64(23.1)
>>>
>>> a.sum(axis=0)
array([5.5, 7.7, 9.9])
>>>
>>> a.sum(axis=1)
array([ 6.6, 16.5])
>>>
>>> np.sum(a)
np.float64(23.1)
>>> a = np.array([[1.1, 2.2, 3.3],
... [4.4, np.nan, 6.6]])
>>>
>>>
>>> a.sum()
np.float64(nan)
>>>
>>> np.sum(a)
np.float64(nan)
>>>
>>> np.nansum(a)
np.float64(17.6)
8.1.3. Cumulative Sum
a.cumsum()np.cumsum()np.nancumsum()
>>> a = np.array([[1.1, 2.2, 3.3],
... [4.4, 5.5, 6.6]])
>>>
>>>
>>> a.cumsum()
array([ 1.1, 3.3, 6.6, 11. , 16.5, 23.1])
>>>
>>> np.cumsum(a)
array([ 1.1, 3.3, 6.6, 11. , 16.5, 23.1])
>>> a = np.array([[1.1, 2.2, 3.3],
... [4.4, np.nan, 6.6]])
>>>
>>>
>>> a.cumsum()
array([ 1.1, 3.3, 6.6, 11. , nan, nan])
>>>
>>> np.cumsum(a)
array([ 1.1, 3.3, 6.6, 11. , nan, nan])
>>>
>>> np.nancumsum(a)
array([ 1.1, 3.3, 6.6, 11. , 11. , 17.6])
8.1.4. Product
a.prod()np.prod()np.nanprod()
>>> a = np.array([[1, 2, 3],
... [4, 5, 6]])
>>>
>>>
>>> a.prod()
np.int64(720)
8.1.5. Cumulative Product
a.cumprod()np.cumprod()np.nancumprod()
8.1.6. Mean
a.mean()np.mean()np.nanmean()
>>> a = np.array([[1, 2, 3],
... [4, 5, 6]])
>>>
>>>
>>> a.mean()
np.float64(3.5)
>>>
>>> a.mean(axis=0)
array([2.5, 3.5, 4.5])
>>>
>>> a.mean(axis=1)
array([2., 5.])
8.1.7. Cumulative Mean
a.cummean()np.cummean()np.nancummean()
8.1.8. Variance
a.var()np.var()np.nanvar()
>>> a = np.array([[1, 2, 3],
... [4, 5, 6]])
>>>
>>>
>>> a.var()
np.float64(2.9166666666666665)
>>>
>>> a.var(axis=0)
array([2.25, 2.25, 2.25])
>>>
>>> a.var(axis=1)
array([0.66666667, 0.66666667])
8.1.9. Standard Deviation
a.std()np.std()np.nanstd()
>>> a = np.array([[1, 2, 3],
... [4, 5, 6]])
>>>
>>>
>>> a.std()
np.float64(1.707825127659933)
>>>
>>> a.std(axis=0)
array([1.5, 1.5, 1.5])
>>>
>>> a.std(axis=1)
array([0.81649658, 0.81649658])
8.1.10. Minimal Value
np.ndarray.argmin()index of annp.ndarray.min()element in arraynp.nanmin()np.nanargmin()
>>> a = np.array([[1, 2, 3],
... [4, 5, 6]])
>>>
>>>
>>> a.min()
np.int64(1)
>>>
>>> a.min(axis=0)
array([1, 2, 3])
>>>
>>> a.min(axis=1)
array([1, 4])
>>> a = np.array([[1, 2, 3],
... [4, 5, 6]])
>>>
>>>
>>> a.argmin()
np.int64(0)
>>>
>>> a.argmin(axis=0)
array([0, 0, 0])
>>>
>>> a.argmin(axis=1)
array([0, 0])
>>> a = np.array([[99, 2, 33],
... [22, 0, 4],
... [4, 155, 6]])
>>>
>>>
>>> a.min()
np.int64(0)
>>>
>>> a.min(axis=0)
array([4, 0, 4])
>>>
>>> a.min(axis=1)
array([2, 0, 4])
>>>
>>> a.min(axis=-1)
array([2, 0, 4])
>>>
>>>
>>> a.argmin()
np.int64(4)
>>>
>>> a.argmin(axis=0)
array([2, 1, 1])
>>>
>>> a.argmin(axis=1)
array([1, 1, 0])
>>>
>>> a.argmin(axis=-1)
array([1, 1, 0])
>>>
>>>
>>> a.flat[4]
np.int64(0)
>>>
>>> np.unravel_index(4, (3, 3))
(np.int64(1), np.int64(1))
>>>
>>> np.unravel_index(a.argmin(), a.shape)
(np.int64(1), np.int64(1))
Shows the coordinates of argmin value:
>>> a = np.array([[1, 2, 3],
... [4, 5, 6]])
>>>
>>>
>>> a.min()
np.int64(1)
>>>
>>> a.argmin()
np.int64(0)
>>>
>>> np.unravel_index(a.argmin(), a.shape)
(np.int64(0), np.int64(0))
>>>
>>> a == a.min()
array([[ True, False, False],
[False, False, False]])
8.1.11. Maximal Value
np.ndarray.argmax()index of ana.max()element in arraynp.nanmax()np.nanargmax()
>>> a = np.array([[1, 2, 3],
... [4, 5, 6]])
>>>
>>>
>>> a.max()
np.int64(6)
>>>
>>> a.max(axis=0)
array([4, 5, 6])
>>>
>>> a.max(axis=1)
array([3, 6])
>>> a = np.array([[1, 2, 3],
... [4, 5, 6]])
>>>
>>>
>>> a.argmax()
np.int64(5)
>>>
>>> a.argmax(axis=1)
array([2, 2])
>>>
>>> a.argmax(axis=0)
array([1, 1, 1])
>>> a = np.array([[99, 2, 33],
... [22, 0, 4],
... [4, 155, 6]])
>>>
>>>
>>> a.max()
np.int64(155)
>>>
>>> a.max(axis=0)
array([ 99, 155, 33])
>>>
>>> a.max(axis=1)
array([ 99, 22, 155])
>>>
>>> a.max(axis=-1)
array([ 99, 22, 155])
>>>
>>>
>>> a.argmax()
np.int64(7)
>>>
>>> a.argmax(axis=0)
array([0, 2, 0])
>>>
>>> a.argmax(axis=1)
array([0, 0, 1])
>>>
>>> a.argmax(axis=-1)
array([0, 0, 1])
>>>
>>> a.flat[7]
np.int64(155)
>>>
>>> np.unravel_index(7, (3, 3))
(np.int64(2), np.int64(1))
>>>
>>> np.unravel_index(a.argmax(), a.shape)
(np.int64(2), np.int64(1))
Shows the coordinates of argmax value:
>>> a = np.array([[1, 2, 3],
... [4, 5, 6]])
>>>
>>>
>>> a.max()
np.int64(6)
>>>
>>> a.argmax()
np.int64(5)
>>>
>>> np.unravel_index(a.argmax(), a.shape)
(np.int64(1), np.int64(2))
>>>
>>> a == a.max()
array([[False, False, False],
[False, False, True]])
8.1.12. Median
np.median()np.nanmedian()
8.1.13. Quantile
np.quantile()np.nanquantile()
8.1.14. Percentile
np.percentile()np.nanpercentile()