NumPy = Numerical Python<\/strong><\/p>\n It is the most important library<\/strong> in the entire Python data\/science\/machine-learning world.<\/p>\n Almost everything serious that happens with numbers in Python uses NumPy under the hood:<\/p>\n If you want to do anything serious with numbers<\/strong> in Python, you must<\/strong> learn NumPy first.<\/p>\n The most important mindset change you need right now:<\/strong><\/p>\n Python lists \u2192 good for general things, shopping lists, names, mixed types \u2192 very slow when doing math on thousands\/millions of numbers<\/p>\n NumPy arrays \u2192 made only for numbers (all elements same type) \u2192 extremely fast mathematical operations \u2192 thinks in whole arrays<\/strong> instead of item-by-item<\/p>\n This single difference makes code 10\u2013100\u00d7 faster<\/strong> and usually much shorter and cleaner.<\/p>\n Almost every person in data science \/ ML \/ scientific computing uses np as the short name. Just accept it \u2014 it\u2019s the universal convention.<\/p>\n Compare with normal Python list:<\/p>\n They look almost the same\u2026 but they are very different<\/strong> inside.<\/p>\n Let\u2019s say we want to give everyone +5 bonus points.<\/p>\n Normal Python way<\/strong> (slow and ugly when list is big):<\/p>\n NumPy way<\/strong> (clean & fast):<\/p>\n No loop!<\/strong> NumPy did the addition on every element automatically<\/strong>.<\/p>\n This is called vectorization<\/strong> \u2014 and it is the #1 reason people love NumPy.<\/p>\n Real-world examples you will see very soon:<\/p>\n All these operations happen element by element<\/strong> \u2014 no loops needed.<\/p>\n Correct ways to copy<\/strong>:<\/p>\n Rule to remember<\/strong>: b = a \u2192 same array (just two names for same data) b = a.copy() \u2192 new independent array<\/p>\n What you should be able to do after this introduction:<\/p>\n Pick one (or tell me what you feel you need most):<\/p>\n Just say a number or describe what feels most useful right now \u2014 we continue from exactly where you are.<\/p>\n You\u2019re doing great \u2014 let\u2019s keep going! \ud83d\ude0a<\/p>\n","protected":false},"excerpt":{"rendered":" NumPy \u2013 The Very First Honest Explanation NumPy = Numerical Python It is the most important library in the entire Python data\/science\/machine-learning world. Almost everything serious that happens with numbers in Python uses NumPy...<\/p>\n","protected":false},"author":1,"featured_media":0,"comment_status":"open","ping_status":"open","sticky":false,"template":"","format":"standard","meta":{"footnotes":""},"categories":[75],"tags":[],"class_list":["post-2464","post","type-post","status-publish","format-standard","hentry","category-numpy"],"_links":{"self":[{"href":"https:\/\/demo.materiamedica.net\/demo6\/wp-json\/wp\/v2\/posts\/2464","targetHints":{"allow":["GET"]}}],"collection":[{"href":"https:\/\/demo.materiamedica.net\/demo6\/wp-json\/wp\/v2\/posts"}],"about":[{"href":"https:\/\/demo.materiamedica.net\/demo6\/wp-json\/wp\/v2\/types\/post"}],"author":[{"embeddable":true,"href":"https:\/\/demo.materiamedica.net\/demo6\/wp-json\/wp\/v2\/users\/1"}],"replies":[{"embeddable":true,"href":"https:\/\/demo.materiamedica.net\/demo6\/wp-json\/wp\/v2\/comments?post=2464"}],"version-history":[{"count":1,"href":"https:\/\/demo.materiamedica.net\/demo6\/wp-json\/wp\/v2\/posts\/2464\/revisions"}],"predecessor-version":[{"id":2465,"href":"https:\/\/demo.materiamedica.net\/demo6\/wp-json\/wp\/v2\/posts\/2464\/revisions\/2465"}],"wp:attachment":[{"href":"https:\/\/demo.materiamedica.net\/demo6\/wp-json\/wp\/v2\/media?parent=2464"}],"wp:term":[{"taxonomy":"category","embeddable":true,"href":"https:\/\/demo.materiamedica.net\/demo6\/wp-json\/wp\/v2\/categories?post=2464"},{"taxonomy":"post_tag","embeddable":true,"href":"https:\/\/demo.materiamedica.net\/demo6\/wp-json\/wp\/v2\/tags?post=2464"}],"curies":[{"name":"wp","href":"https:\/\/api.w.org\/{rel}","templated":true}]}}\n
Let\u2019s Start \u2013 First Code You Should Type<\/h3>\n
import numpy as np<\/code><\/pre>\n<\/div>\n<\/div>\n<\/div>\n<\/div>\n# Let's create our very first NumPy array\r\nscores = np.array([78, 92, 65, 84, 71, 88])\r\nprint(scores)\r\n# output: [78 92 65 84 71 88]<\/code><\/pre>\n<\/div>\n<\/div>\n<\/div>\n<\/div>\npython_list = [78, 92, 65, 84, 71, 88]\r\nprint(python_list)\r\n# output: [78, 92, 65, 84, 71, 88]<\/code><\/pre>\n<\/div>\n<\/div>\n<\/div>\n<\/div>\nWhy NumPy Feels So Different \u2013 First Magic Example<\/h3>\n
new_scores = []\r\nfor score in python_list:\r\n new_scores.append(score + 5)\r\nprint(new_scores)<\/code><\/pre>\n<\/div>\n<\/div>\n<\/div>\n<\/div>\nbonus = scores + 5\r\nprint(bonus)\r\n# [83 97 70 89 76 93]<\/code><\/pre>\n<\/div>\n<\/div>\n<\/div>\n<\/div>\nMost Common First Arrays You Will Create<\/h3>\n
# 1. From a list (most common way at beginning)\r\na = np.array([1.5, 2.7, 3.2, 4.0])\r\n\r\n# 2. All zeros (very useful to initialize)\r\nzeros = np.zeros(10) # 1D\r\nzeros2d = np.zeros((4, 6)) # 4 rows, 6 columns\r\n\r\n# 3. All ones\r\nones = np.ones((3, 5))\r\n\r\n# 4. Fill with same number\r\nprices = np.full((2, 8), 99.99)\r\n\r\n# 5. Sequence of numbers (very common!)\r\nar = np.arange(0, 20, 2) # 0,2,4,...,18\r\nlin = np.linspace(0, 1, 11) # 0.0, 0.1, ..., 1.0\r\n\r\n# 6. Random numbers \u2013 you will use these A LOT\r\nnp.random.seed(42) # makes random results repeatable\r\n\r\nuniform = np.random.rand(5) # 5 random numbers between 0 and 1\r\nnormal = np.random.randn(1000) # 1000 numbers from normal distribution\r\ndice = np.random.randint(1, 7, size=20) # 20 dice rolls<\/code><\/pre>\n<\/div>\n<\/div>\n<\/div>\n<\/div>\nThe 4 Most Important Properties \u2013 Check These Every Time<\/h3>\n
data = np.random.randint(0, 100, size=(3, 4))\r\n\r\nprint(data)\r\n# example output:\r\n# [[45 68 37 91]\r\n# [12 73 8 29]\r\n# [64 19 83 52]]\r\n\r\nprint(data.shape) # (3, 4) \u2190 MOST IMPORTANT LINE\r\nprint(data.ndim) # 2 \u2190 how many dimensions\r\nprint(data.size) # 12 \u2190 total numbers\r\nprint(data.dtype) # int64 \u2190 what type are the numbers<\/code><\/pre>\n<\/div>\n<\/div>\n<\/div>\n<\/div>\n# Machine learning dataset\r\nX.shape \u2192 (number_of_samples, number_of_features) # e.g. (10000, 784)\r\n\r\n# Images (very common)\r\nimage.shape \u2192 (height, width, 3) # RGB image\r\nbatch_images.shape \u2192 (32, 224, 224, 3) # 32 images of 224\u00d7224 pixels<\/code><\/pre>\n<\/div>\n<\/div>\n<\/div>\n<\/div>\nFirst Useful Math You Can Do Immediately<\/h3>\n
temps = np.array([23.4, 25.1, 19.8, 28.7, 22.0])\r\n\r\nprint(temps * 1.8 + 32) # convert \u00b0C \u2192 \u00b0F\r\n# [74.12 77.18 67.64 83.66 71.6 ]\r\n\r\nprint(temps ** 2) # square each value\r\nprint(np.sqrt(temps)) # square root\r\nprint(np.round(temps)) # round to nearest integer\r\nprint(temps > 24) # boolean array: True\/False<\/code><\/pre>\n<\/div>\n<\/div>\n<\/div>\n<\/div>\nVery Common Beginner Mistake #1 \u2013 Copying<\/h3>\n
a = np.array([10, 20, 30, 40])\r\nb = a # \u2190 DANGER! This is NOT a real copy\r\n\r\nb[0] = 999\r\nprint(a) # [999 20 30 40] \u2190 a also changed!<\/code><\/pre>\n<\/div>\n<\/div>\n<\/div>\n<\/div>\nc = a.copy() # safest and clearest\r\nd = np.copy(a)\r\ne = a[:] # usually works, but be careful with slices<\/code><\/pre>\n<\/div>\n<\/div>\n<\/div>\n<\/div>\nQuick Summary \u2013 Your First NumPy Survival Kit<\/h3>\n
\n
Where should we go next?<\/h3>\n
\n