NumPy: Vectorize your brain

NumPy is the fundamental Python package for scientific computing. However, being efficient with NumPy might require slightly changing how you write Python code.

I’m going to show you the basic idioms essential for fast numerical computations in Python with NumPy. We’ll see why Python loops are slow and why vectorizing these operations with NumPy can often be good.

Transcript

1. Numpy: Vectorize your brain Ekaterina Tuzova
2. K nearest neighbors https://archive.ics.uci.edu/ml/datasets/Wine
3. NumPy
4. What is NumPy? Numpy is the fundamental package for scientific computing with Python.
5. IPython
6. Python and Performance
7. Python is fast
8. Python is slow
9. Euclidian distance
10. “Magic” timeit
11. Euclidian distance. C
12. Euclidian distance. C
13. Euclidian distance
14. line_profiler and “magic” lprun
15. Euclidian distance
16. Compiled languages
17. Interpreted languages
18. What can be done?
19. NumPy
20. Ufuncs
21. Universal functions Special type of function defined within a numpy library and it operate element-wise on arrays.
22. Arithmetic operations
23. Arithmetic operations
24. Arithmetic operations
25. Arithmetic operations
26. Ufuncs available - Arithmetic - Bitwise - Comparison - Trigonometric - Floating …
27. Slicing and indexing
28. Slicing and indexing
29. Slicing and indexing
30. Multidimensional arrays
31. Multidimensional arrays
32. Index arrays
33. Index arrays
34. Index arrays
35. Masking
36. Masking
37. Test train split
38. Test train split
39. Broadcasting
40. Broadcasting Broadcasting describes how NumPy treats arrays with different shapes during arithmetic operations.
41. Broadcasting rules 1.  If two arrays differ in their number of dimension, the shape of the array with the fewer dimensions is padded with ones on it’s leading(left) size. 2.  If the shape of two arrays doesn’t match in any dimension, the array with shape equal to 1 in that dimension is stretched to match the other shape. 3.  If these conditions are not met, raise a ValueError: operands could not be broadcast together with shapes
42. Broadcasting. Example
43. np.newaxis
44. np.newaxis
45. np.newaxis
46. Aggregations
47. Aggregations
48. Aggregations
49. NumPy resume Basic ideas to make you code faster: - Ufuncs - Slicing and indexing - Broadcasting - Aggregations
50. k-means
51. Algorithm 1.  Select k points at random as cluster centers. 2.  Assign objects to their closest cluster center according to the Euclidean distance function. 3.  Calculate the centroid or mean of all objects in each cluster. 4.  Repeat steps 2, 3 and 4 until the same points are assigned to each cluster in consecutive rounds.
52. Synthetic data
54. Vectorized euclidian distance
55. k-means
57. Thank you. @ktisha