<p>You most probably will know by now that data storytelling, accomplished by data visualization, amongst other things, is an essential skill for every data scientist: after you have turned the raw data into understanding, insights and knowledge, you also need to communicate these findings effectively to your audience. </p>
<p>For most beginners, the first Python data visualization library that they use is, naturally, Matplotlib. It is a Python 2D plotting library that enables users to make publication-quality figures. It is quite an extensive library where a <a href="https://www.datacamp.com/community/blog/python-matplotlib-cheat-sheet">cheat sheet</a> will definitely come in handy when you&#39;re learning, but when you manage to use this library effectively, you&#39;ll also be able to get insights and work better with other packages, such as Pandas, that intend to build more plotting integration with Matplotlib as time goes on. </p>
<p>Another package that you&#39;ll be able to tackle easily is Seaborn, the statistical data visualization library of Python. </p>
<p>DataCamp has created a Seaborn cheat sheet for those who are ready to get started with this data visualization library with the help of a handy one-page reference. </p>
<p>You&#39;ll see that this cheat sheet presents you with the five basic steps that you can go through to make beautiful statistical graphs in Python. </p>
<p>Check out the infographic by clicking on the button below:</p>
<p><a href="http://datacamp-community-prod.s3.amazonaws.com/263130e2-2c92-4348-a356-9ed9b5034247"><img alt="Seaborn Python cheat sheet" src="https://res.cloudinary.com/dyd911kmh/image/upload/f_auto,q_auto:best/v1625839419/Python_Seaborn_Cheat_Sheet_q074wv.png" style="width: 800px; height: 419px;" /></a></p>
<p>This cheat sheet will walk you through the five steps that you need to go through to make these plots: you&#39;ll see how you can load in data, set the figure aesthetics, plot, customize and eventually, show or save your plot with Seaborn.</p>
<p>What might have looked difficult before will definitely be more clear once you start using this cheat sheet! Use it in combination with the <a href="http://seaborn.pydata.org/examples/index.html">Seaborn Gallery</a>, the <a href="http://seaborn.pydata.org/index.html">documentation</a>, and our <a href="https://www.datacamp.com/community/tutorials/seaborn-python-tutorial">Seaborn tutorial</a>.</p>
<p>Also, don&#39;t miss out on our other cheat sheets for data science that cover <a href="https://www.datacamp.com/community/blog/python-scipy-cheat-sheet/">SciPy</a>, <a href="https://www.datacamp.com/community/blog/python-numpy-cheat-sheet/">NumPy</a>, <a href="https://www.datacamp.com/community/blog/scikit-learn-cheat-sheet/">Scikit-Learn</a>, <a href="https://www.datacamp.com/community/blog/bokeh-cheat-sheet-python/">Bokeh</a>, <a href="https://www.datacamp.com/community/blog/python-pandas-cheat-sheet/">Pandas</a> and the <a href="https://www.datacamp.com/community/tutorials/python-data-science-cheat-sheet-basics/">Python basics</a>.</p>


https://www.datacamp.com/community/blog/seaborn-cheat-sheet-python

This Python Seaborn cheat sheet with code samples guides you through the data visualization library that is based on Matplotlib.

This Python Seaborn cheat sheet with code samples guides you through the data visualization library that is based on matplotlib.

Python Seaborn Cheat Sheet

<p>You&#39;ll probably know the Jupyter notebooks pretty well - it&#39;s one of the most well-known parts of the Jupyter ecosystem! If you haven&#39;t explored the ecosystem yet or if you simply want to know more about it, don&#39;t hesitate to go and explore it <a href="www.jupyter.org">here</a>!.</p>
<p>For those who are new to Project Jupyter, the Jupyter Notebook Application produces documents that contain a mix of executable code, text elements, and even HTML, which makes it thé ideal place to bring together an analysis description and its results as well as to perform data analysis in real time. This, combined with its many useful functionalities, explains why the Jupyter Notebook is a <a href="https://www.datacamp.com/community/tutorials/data-science-python-ide">one of the data scientist&#39;s preferred development environments environments</a> that allows for interactive, reproducible data science analysis, computation and communication.</p>
<p>One of the other great things about the Jupyter notebooks?</p>
<p>They&#39;re super easy to get started with! You might have already noticed this when you read DataCamp&#39;s <a href="https://www.datacamp.com/community/tutorials/tutorial-jupyter-notebook">definitive guide to Jupyter Notebook</a>. However, when you first enter in the application, you might have to find your way around the variety of functionalities that are presented to you: from saving your current notebook to adding or moving cells in the notebook or embedding current widgets in your notebook - without a doubt, there&#39;s a lot out there to discover when you first get started! </p>
<p>That&#39;s why DataCamp made a Jupyter Notebook cheat sheet for those who are just starting out and that want to have some help to find their way around. </p>
<p><strong>Note</strong> also that the Jupyter Notebook Application has a handy &quot;Help&quot; menu that includes a full-blown User Interface Tour! -- No worries, we also included this in the cheat sheet :)</p>
<p>Check it out here:</p>
<p><a href="http://datacamp-community-prod.s3.amazonaws.com/21fdc814-3f08-4aa9-90fa-247eedefd655" target="_blank"><img src="https://res.cloudinary.com/dyd911kmh/image/upload/f_auto,q_auto:best/v1625839536/Jupyter_Notebook_Cheat_Sheet_ntmalw.png" alt="Jupyter Notebook cheat sheet"/></a></p>
<p>In short, this cheat sheet will help you to kickstart your data science projects, however small or big they might be: with of some screenshots and explanations, you&#39;ll be a Jupyter Notebook expert in no time!</p>
<p>So what are you waiting for? Time to get started!</p>


https://www.datacamp.com/community/blog/jupyter-notebook-cheat-sheet

This Jupyter Notebook cheat sheet will help you to find your way around the well-known Notebook App, a subproject of Project Jupyter.

This Jupyter Notebook cheat sheet will help you to find your way around the well-known Jupyter Notebook App, a subproject of Project Jupyter.

Jupyter Notebook Cheat Sheet

<p><a href="http://keras.io/">Keras</a> is an easy-to-use and powerful library for Theano and TensorFlow that provides a high-level neural networks API to develop and evaluate deep learning models.</p>

<p>We recently launched one of the first online interactive deep learning course using Keras 2.0, called &quot;<a href="https://www.datacamp.com/courses/deep-learning-in-python/">Deep Learning in Python</a>&quot;.</p>

<p>Now, DataCamp has created a Keras cheat sheet for those who have already taken the course and that still want a handy one-page reference or for those who need an extra push to get started.</p>

<p>In no time, this Keras cheat sheet will make you familiar with how you can load datasets from the library itself, preprocess the data, build up a model architecture, and compile, train, and evaluate it. As there is a considerable amount of freedom in how you build up your models, you&#39;ll see that the cheat sheet uses some of the simple&nbsp;key&nbsp;code examples of the Keras library that you need to know to get started with building your own neural networks in Python.</p>

<p>Furthermore, you&#39;ll also see some examples of how to inspect your model, and how you can save and reload it. Lastly, you&rsquo;ll also find examples of how you can predict values for test data and how you can fine tune your models by adjusting the optimization parameters and early stopping.</p>

<p>In short, you&#39;ll see that this cheat sheet not only presents you with the six steps that you can go through to make neural networks in Python with the Keras library.</p>

<p><a href="http://datacamp-community-prod.s3.amazonaws.com/af9bb467-170d-41c9-a0bd-26e675384c4e" target="_blank"><img alt="keras cheat sheet" src="https://res.cloudinary.com/dyd911kmh/image/upload/f_auto,q_auto:best/v1625839502/Keras_Cheat_Sheet-_Neural_Networks_in_Python_tjhi1q.png" style="width: 800px; height: 551px;" /></a></p>

<p>In short, this cheat sheat will boost your journey with deep learning in Python: you&#39;ll have preprocessed, created, validated and tuned your deep learning models in no time thanks to the code examples! &nbsp; &nbsp;</p>

<p><strong>(Click above to download a printable version or read the online version below.) </strong></p>

<div id="scoped-content">
<style type="text/css">:target:before { content:""; display:block; height:150px; margin:-150px 0 0; } h3 {font-weight:normal; } h4 { 	font-weight: lighter; }
</style>
<h2>Python For Data Science Cheat Sheet: Keras</h2>

<p>Keras is a powerful and easy-to-use deep learning library for Theano and TensorFlow that provides a high-level neural networks API to develop and evaluate deep learning models.</p>

<h3>A Basic Example</h3>

<pre>
<code>&gt;&gt;&gt; import numpy as np
&gt;&gt;&gt; from keras.models import Sequential
&gt;&gt;&gt; from keras.layers import Dense
&gt;&gt;&gt; data = np.random.random((1000,100))
&gt;&gt;&gt; labels = np.random.randint(2,size=(1000,1))
&gt;&gt;&gt; model = Sequential()
&gt;&gt;&gt; model.add(Dense(32, activation='relu', input_dim=100))
&gt;&gt;&gt; model.add(Dense(1, activation='sigmoid'))
&gt;&gt;&gt; model.compile(optimizer='rmsprop', loss='binary_crossentropy', metrics=['accuracy'])
&gt;&gt;&gt;  model.fit(data,labels,epochs=10,batch_size=32)
&gt;&gt;&gt;  predictions = model.predict(data)</code></pre>

<h3>Data</h3>

<p>Your data needs to be stored as NumPy arrays or as a list of NumPy arrays. Ideally, you split the data in training and test sets, for which you can also resort to the <code>train_test_split</code> module of <code>sklearn.cross_validation</code>.</p>

<h4>Keras Data Sets</h4>

<pre>
<code>&gt;&gt;&gt; from keras.datasets import boston_housing, mnist, cifar10, imdb
&gt;&gt;&gt; (x_train,y_train),(x_test,y_test) = mnist.load_data()
&gt;&gt;&gt; (x_train2,y_train2),(x_test2,y_test2) = boston_housing.load_data()
&gt;&gt;&gt; (x_train3,y_train3),(x_test3,y_test3) = cifar10.load_data()
&gt;&gt;&gt; (x_train4,y_train4),(x_test4,y_test4) = imdb.load_data(num_words=20000)
&gt;&gt;&gt; num_classes = 10</code></pre>

<h4>Other</h4>

<pre>
<code>&gt;&gt;&gt; from urllib.request import urlopen
&gt;&gt;&gt; data = np.loadtxt(urlopen("http://archive.ics.uci.edu/ml/machine-learning-databases/pima-indians-diabetes/pima-indians-diabetes.data"),delimiter=",")
&gt;&gt;&gt; X = data[:,0:8]
&gt;&gt;&gt; y = data [:,8]</code></pre>

<h3>Preprocessing</h3>

<h4>Sequence Padding</h4>

<pre>
<code>&gt;&gt;&gt; from keras.preprocessing import sequence
&gt;&gt;&gt; x_train4 = sequence.pad_sequences(x_train4,maxlen=80)
&gt;&gt;&gt; x_test4 = sequence.pad_sequences(x_test4,maxlen=80)</code></pre>

<h4>One-Hot Encoding</h4>

<pre>
<code>&gt;&gt;&gt; from keras.utils import to_categorical
&gt;&gt;&gt; Y_train = to_categorical(y_train, num_classes)
&gt;&gt;&gt; Y_test = to_categorical(y_test, num_classes)
&gt;&gt;&gt; Y_train3 = to_categorical(y_train3, num_classes)
&gt;&gt;&gt; Y_test3 = to_categorical(y_test3, num_classes)</code></pre>

<h4>Train And Test Sets</h4>

<pre>
<code>&gt;&gt;&gt; from sklearn.model_selection import train_test_split
&gt;&gt;&gt; X_train5, X_test5, y_train5, y_test5 = train_test_split(X, y, test_size=0.33, random_state=42)</code></pre>

<h3>Standardization/Normalization</h3>

<pre>
<code>&gt;&gt;&gt; from sklearn.preprocessing import StandardScaler
&gt;&gt;&gt; scaler = StandardScaler().fit(x_train2)
&gt;&gt;&gt; standardized_X = scaler.transform(x_train2)
&gt;&gt;&gt; standardized_X_test = scaler.transform(x_test2)</code></pre>

<h3>Model Architecture</h3>

<h4>Sequential Model</h4>

<pre>
<code>&gt;&gt;&gt; from keras.models import Sequential
&gt;&gt;&gt; model = Sequential()
&gt;&gt;&gt; model2 = Sequential()
&gt;&gt;&gt; model3 = Sequential()</code></pre>

<h4>Multi-Layer Perceptron (MLP)</h4>

<p><strong>Binary Classification</strong></p>

<pre>
<code>&gt;&gt;&gt; from keras.layers import Dense
&gt;&gt;&gt; model.add(Dense(12, input_dim=8, kernel_initializer='uniform', activation='relu'))
&gt;&gt;&gt; model.add(Dense(8, kernel_initializer='uniform', activation='relu'))
&gt;&gt;&gt; model.add(Dense(1, kernel_initializer='uniform', activation='sigmoid'))</code></pre>

<p><strong>Multi-Class Classification</strong></p>

<pre>
<code>&gt;&gt;&gt; from keras.layers import Dropout
&gt;&gt;&gt; model.add(Dense(512,activation='relu',input_shape=(784,)))
&gt;&gt;&gt; model.add(Dropout(0.2))
&gt;&gt;&gt; model.add(Dense(512,activation='relu'))
&gt;&gt;&gt; model.add(Dropout(0.2))
&gt;&gt;&gt; model.add(Dense(10,activation='softmax'))</code></pre>

<p><strong>Regression</strong></p>

<pre>
<code>&gt;&gt;&gt; model.add(Dense(64, activation='relu', input_dim=train_data.shape[1]))
&gt;&gt;&gt; model.add(Dense(1))</code></pre>

<h4>Convolutional Neural Network (CNN)</h4>

<pre>
<code>&gt;&gt;&gt; from keras.layers import Activation, Conv2D, MaxPooling2D, Flatten
&gt;&gt;&gt; model2.add(Conv2D(32, (3,3), padding='same', input_shape=x_train.shape[1:]))
&gt;&gt;&gt; model2.add(Activation('relu'))
&gt;&gt;&gt; model2.add(Conv2D(32, (3,3)))
&gt;&gt;&gt; model2.add(Activation('relu'))
&gt;&gt;&gt; model2.add(MaxPooling2D(pool_size=(2,2)))
&gt;&gt;&gt; model2.add(Dropout(0.25))
&gt;&gt;&gt; model2.add(Conv2D(64, (3,3), padding='same'))
&gt;&gt;&gt; model2.add(Activation('relu'))
&gt;&gt;&gt; model2.add(Conv2D(64, (3, 3)))
&gt;&gt;&gt; model2.add(Activation('relu'))
&gt;&gt;&gt; model2.add(MaxPooling2D(pool_size=(2,2)))
&gt;&gt;&gt; model2.add(Dropout(0.25))
&gt;&gt;&gt; model2.add(Flatten())
&gt;&gt;&gt; model2.add(Dense(512))
&gt;&gt;&gt; model2.add(Activation('relu'))
&gt;&gt;&gt; model2.add(Dropout(0.5))
&gt;&gt;&gt; model2.add(Dense(num_classes))
&gt;&gt;&gt; model2.add(Activation('softmax'))</code></pre>

<h4>Recurrent Neural Network (RNN)</h4>

<pre>
<code>&gt;&gt;&gt; from keras.klayers import Embedding,LSTM
&gt;&gt;&gt; model3.add(Embedding(20000,128))
&gt;&gt;&gt; model3.add(LSTM(128,dropout=0.2,recurrent_dropout=0.2))
&gt;&gt;&gt; model3.add(Dense(1,activation='sigmoid'))</code></pre>

<h3>Inspect Model</h3>

<p>Model output shape</p>

<pre>
<code>&gt;&gt;&gt; model.output_shape</code></pre>

<p>Model summary representation</p>

<pre>
<code>&gt;&gt;&gt; model.summary()</code></pre>

<p>Model configuration</p>

<pre>
<code>&gt;&gt;&gt; model.get_config()</code></pre>

<p>List all weight tensors in the model</p>

<pre>
<code>&gt;&gt;&gt; model.get_weights()</code></pre>

<h3>Compile Model</h3>

<h4>Multi-Layer Perceptron (MLP)</h4>

<p><strong>MLP: Binary Classification</strong></p>

<pre>
<code>&gt;&gt;&gt; model.compile(optimizer='adam', loss='binary_crossentropy', metrics=['accuracy'])</code></pre>

<p><strong>MLP: Multi-Class Classification</strong></p>

<pre>
<code>&gt;&gt;&gt; model.compile(optimizer='rmsprop', loss='categorical_crossentropy', metrics=['accuracy'])</code></pre>

<p><strong>MLP: Regression</strong></p>

<pre>
<code>&gt;&gt;&gt; model.compile(optimizer='rmsprop', loss='mse', metrics=['mae'])</code></pre>

<h4>Recurrent Neural Network (RNN)</h4>

<pre>
<code>&gt;&gt;&gt; model3.compile(loss='binary_crossentropy', optimizer='adam', metrics=['accuracy'])</code></pre>

<h3>Model Training</h3>

<pre>
<code>&gt;&gt;&gt; model3.fit(x_train4, y_train4, batch_size=32, epochs=15, verbose=1, validation_data=(x_test4, y_test4))</code></pre>

<h3>Evaluate Your Model&#39;s Performance</h3>

<pre>
<code>&gt;&gt;&gt; score = model3.evaluate(x_test, y_test, batch_size=32)</code></pre>

<h3>Prediction</h3>

<pre>
<code>&gt;&gt;&gt; model3.predict(x_test4, batch_size=32)
&gt;&gt;&gt; model3.predict_classes(x_test4,batch_size=32)</code></pre>

<h3>Save/Reload Models</h3>

<pre>
<code>&gt;&gt;&gt; from keras.models import load_model
&gt;&gt;&gt; model3.save('model_file.h5')
&gt;&gt;&gt; my_model = load_model('my_model.h5')</code></pre>

<h3>Model Fine-Tuning</h3>

<h4>Optimization Parameters</h4>

<pre>
<code>&gt;&gt;&gt; from keras.optimizers import RMSprop
&gt;&gt;&gt; opt = RMSprop(lr=0.0001, decay=1e-6)
&gt;&gt;&gt; model2.compile(loss='categorical_crossentropy', optimizer=opt, metrics=['accuracy'])</code></pre>

<h4>Early Stopping</h4>

<pre>
<code>&gt;&gt;&gt; from keras.callbacks import EarlyStopping
&gt;&gt;&gt; early_stopping_monitor = EarlyStopping(patience=2)
&gt;&gt;&gt; model3.fit(x_train4, y_train4, batch_size=32, epochs=15, validation_data=(x_test4, y_test4), callbacks=[early_stopping_monitor])</code></pre>

<h3>Going Further</h3>

<p>Begin with <a href="https://www.datacamp.com/community/tutorials/deep-learning-python">our Keras tutorial for beginners</a>, in which you&#39;ll learn in an easy, step-by-step way how to explore and preprocess the wine quality data set, build up a multi-layer perceptron for classification and regression tasks, compile, fit and evaluate the model and fine-tune the model that you have built.</p>

<p>Also, don&#39;t miss out on our<a href="https://www.datacamp.com/community/blog/scikit-learn-cheat-sheet/"> Scikit-Learn cheat sheet</a>, <a href="https://www.datacamp.com/community/blog/python-numpy-cheat-sheet/">NumPy cheat sheet</a> and <a href="https://www.datacamp.com/community/blog/python-pandas-cheat-sheet/">Pandas cheat sheet</a>!</p>
</div>


https://www.datacamp.com/community/blog/keras-cheat-sheet

Make your own neural networks with this Keras cheat sheet to deep learning in Python for beginners, with code samples. 

Keras Cheat Sheet: Neural Networks in Python

<p>DataCamp Digest is our newsletter aimed at providing the most up-to-date insights and news on all things data science. In this edition of the newsletter, we discuss the dark side of algorithmic hiring, the rise of AI sandboxes for compliant innovation, our future with AI, and more.</p>
<p><center><a target="_blank" href="https://www.datacamp.com/groups/business"><img src="https://res.cloudinary.com/dyd911kmh/image/upload/f_auto,q_auto:best/v1612888906/insights_qhc4kg.png" /></a></center></p>

<h4 id="-will-ai-give-us-the-lives-of-leisure-we-long-for-https-www-nytimes-com-2021-06-11-opinion-ezra-klein-podcast-sam-altman-html-podcast-the-ezra-klein-show"><a target="_blank" href="https://www.nytimes.com/2021/06/11/opinion/ezra-klein-podcast-sam-altman.html">Will AI give us the lives of leisure we long for?</a> | Podcast: The Ezra Klein Show</h4>
<p>Sam Altman is the CEO of OpenAI, the research lab that developed and rolled out GPT-3. Sam is a big believer in Moore’s Law, which posits that computing power doubles every two years despite shrinking computer costs.. Altman predicts that the advent of highly intelligent agents foreshadowed by GPT-3, will generate a “Moore’s Law for everything”, unlocking greater quality and more affordable housing, health care, education, you name it. Listen to the show to learn more about Altman views on the future of AI.</p>
<h4 id="-to-regulate-ai-try-playing-in-a-sandbox-https-www-morningbrew-com-emerging-tech-stories-2021-05-26-regulate-ai-just-play-sandbox-emerging-tech-brew"><a target="_blank" href="https://www.morningbrew.com/emerging-tech/stories/2021/05/26/regulate-ai-just-play-sandbox">To regulate AI, try playing in a sandbox</a> | Emerging Tech Brew</h4>
<p>There is rising interest in using “regulatory sandboxes” to govern AI without hamstringing innovation. These allow organizations to develop and test new technologies in a low-stakes, monitored environment before rolling them out to the general public. How will this change the AI landscape? Read more to find out.</p>
<h4 id="-what-google-s-ai-designed-chips-tells-us-about-the-nature-of-intelligence-https-thenextweb-com-news-what-google-ai-designed-chip-tells-nature-intelligence-syndication-tnw-neural"><a target="_blank" href="https://thenextweb.com/news/what-google-ai-designed-chip-tells-nature-intelligence-syndication">What Google’s AI-designed chips tells us about the nature of intelligence</a> | TNW Neural</h4>
<p>Google’s AI research team built a reinforcement learning model that helps the company design more efficient chips. As the components in a chip grow, finding the most efficient location for each component is a major challenge. Learn how Google leveraged human and artificial intelligence  to solve this problem.</p>
<p><br><center><a target="_blank" href="https://www.datacamp.com/groups/business"><img src="https://res.cloudinary.com/dyd911kmh/image/upload/f_auto,q_auto:best/v1612888907/Editor_s_Picks_nj17hh.png" /></a></center></p>

<h4 id="-hired-by-an-algorithm-https-www-technologyreview-com-2021-06-23-1026918-podcast-hired-by-an-algorithm-mit-technology-review"><a target="_blank" href="https://www.technologyreview.com/2021/06/23/1026918/podcast-hired-by-an-algorithm/">Hired by an algorithm</a> | MIT Technology Review</h4>
<p>In this podcast, the CEO of ZipRecruiter and one of LinkedIn’s architects behind its algorithmic job-matching system discussed the advantages and disadvantages of algorithmic recruitment practices. Even though software can accelerate the recruitment processes, algorithms could have biased results  based on race, gender, and in at least one case, whether you played lacrosse in high school.</p>
<h4 id="-ai-can-now-emulate-text-style-in-images-using-just-a-single-word-https-ai-facebook-com-blog-ai-can-now-emulate-text-style-in-images-in-one-shot-using-just-a-single-word-facebook-ai"><a target="_blank" href="https://ai.facebook.com/blog/ai-can-now-emulate-text-style-in-images-in-one-shot-using-just-a-single-word">AI can now emulate text style in images using just a single word</a> | Facebook AI</h4>
<p>The team at Facebook AI developed a self-supervised learning algorithm that can emulate the style of a text in a photo using just one training example. Soon enough, editing highly stylized or handwritten text in images will be as simple as copy paste.</p>
<h4 id="-global-ai-vibrancy-tool-https-aiindex-stanford-edu-vibrancy-stanford-university-human-centered-ai"><a target="_blank" href="https://aiindex.stanford.edu/vibrancy/">Global AI Vibrancy Tool</a> | Stanford University Human Centered AI</h4>
<p>The Global AI Vibrancy Tool is an interactive visualization that allows cross-country comparison for up to 26 countries across 22 indicators. The tool provides transparent evaluation of the relative position of countries based on wide-ranging categories such as research, economy and inclusion.</p>
<p><br><center><a target="_blank" href="https://www.datacamp.com/groups/business"><img src="https://res.cloudinary.com/dyd911kmh/image/upload/f_auto,q_auto:best/v1612889346/Inspiration2_wxr25i.png" /></a></center></p>

<h4 id="-how-airbnb-standardized-metric-computation-at-scale-part-i-https-medium-com-airbnb-engineering-how-airbnb-achieved-metric-consistency-at-scale-f23cc53dea70-the-airbnb-tech-blog"><a target="_blank" href="https://medium.com/airbnb-engineering/how-airbnb-achieved-metric-consistency-at-scale-f23cc53dea70">How Airbnb standardized metric computation at scale Part I</a> | The Airbnb Tech Blog</h4>
<p>As data warehouses keep expanding, the challenge of creating, managing, computing, and distributing data across different departments and teams keeps growing. Learn why Airbnb built Minerva and how the software allowed them to turn data into actionable strategies. Make sure you read Part II of the story.</p>
<h4 id="-9-ethical-ai-principles-for-organizations-to-follow-https-www-weforum-org-agenda-2021-06-ethical-principles-for-ai-world-economic-forum"><a target="_blank" href="https://www.weforum.org/agenda/2021/06/ethical-principles-for-ai">9 Ethical AI Principles for Organizations to Follow</a> | World Economic Forum</h4>
<p>An increasing number of organizations are beginning to craft ethical charters and principles that will guide their AI development. In this article, the responsible AI team at PwC lays out 9 ethical principles for developing AI that can be leveraged by any organization today.</p>
<h4 id="-exploring-data-at-netflix-https-netflixtechblog-com-exploring-data-netflix-9d87e20072e3-the-netflix-tech-blog"><a target="_blank" href="https://netflixtechblog.com/exploring-data-netflix-9d87e20072e3">Exploring Data at Netflix</a> | The Netflix Tech Blog</h4>
<p>Providing easy access to relevant data to a variety of data roles across an organization is no easy feat. In this article, the team at Netflix outlines how they scaled access to data with their internal Data Explorer tool, and how they’re open-sourcing it for the world to use.</p>
<p><br><center><a target="_blank" href="https://www.datacamp.com/groups/business"><img src="https://res.cloudinary.com/dyd911kmh/image/upload/f_auto,q_auto:best/v1612888937/Tools_and_Techniques_sf5hut.png" /></a></center></p>

<h4 id="-ivy-the-templated-deep-learning-framework-https-ivy-dl-org-deep-learning-weekly"><a target="_blank" href="https://ivy-dl.org/">IVY: The templated deep learning framework</a> | Deep Learning Weekly</h4>
<p>Ivy is a templated deep learning framework which maximizes the portability of deep learning codebases by wrapping the functional APIs of existing frameworks. It currently supports Jax, TensorFlow, PyTorch, MXNet and Numpy.</p>
<h4 id="-orbit-an-open-source-package-for-time-series-interference-and-forecasting-https-eng-uber-com-orbit-uber-engineering"><a target="_blank" href="https://eng.uber.com/orbit/">Orbit, an open source package for time series interference and forecasting</a> | Uber Engineering</h4>
<p>Orbit is a newly developed package by the team at Uber for Bayesian time series modeling. The goal behind Orbit is to create a tool that is easy to use, flexible, interitible, and high performing while allowing for easy model specification and analysis but not limiting itself to a small subset of models.</p>
<h4 id="-gradio-create-quick-uis-for-prototyping-your-ml-model-https-www-gradio-app-the-gradio-team"><a target="_blank" href="https://www.gradio.app/">Gradio: Create quick UIs for prototyping your ML model</a> | The Gradio Team</h4>
<p>Creating UIs for testing machine learning models can be time consuming sometimes. Gradio is an open source tool that lets you create simple interfaces to demo your models.</p>
<p><br><center><a target="_blank" href="https://www.datacamp.com/groups/business"><img src="https://res.cloudinary.com/dyd911kmh/image/upload/f_auto,q_auto:best/v1612889508/From_DataCamp_c6kasv.png" /></a></center></p>

<h4 id="-webinar-developing-an-ai-literate-nation-https-www-datacamp-com-resources-webinars-developing-an-ai-literate-nation-"><a target="_blank" href="https://www.datacamp.com/resources/webinars/developing-an-ai-literate-nation">Webinar: Developing an AI literate nation</a></h4>
<p>In this fireside chat, Laurence Liew, Director of AI Innovation, and Koo Sengmeng, Senior Deputy Director of AI Innovation at AI Singapore will deep dive into AI Singapore’s mission and how it is accelerating AI Adoption across the nation.</p>
<h4 id="-white-paper-data-literacy-for-responsible-ai-https-www-datacamp-com-resources-whitepapers-data-literacy-for-responsible-ai-"><a target="_blank" href="https://www.datacamp.com/resources/whitepapers/data-literacy-for-responsible-ai">White Paper: Data literacy for responsible AI</a></h4>
<p>In this white paper, co-written with the trusted AI team at DataRobot we outline the importance of developing responsible AI, practical solutions data teams and organizations can adopt, and the crucial role data literacy plays when scaling responsible AI.</p>
<h4 id="-podcast-64-creating-trust-in-data-with-data-observability-https-www-datacamp-com-community-podcast-creating-trust-in-data-with-data-observability-"><a target="_blank" href="https://www.datacamp.com/community/podcast/creating-trust-in-data-with-data-observability">Podcast: #64 Creating trust in data with data observability</a></h4>
<p>In this episode of DataFramed, Adel speaks with Barr Moses, CEO, and co-founder of Monte Carlo on the importance of data quality and how data teams can leverage data observability to create trust in data. Make sure to listen and subscribe on your favorite podcasting app.</p>
<p><br><center><a target="_blank" href="https://www.datacamp.com/business/demo"><img src="https://res.cloudinary.com/dyd911kmh/image/upload/f_auto,q_auto:best/v1603718763/banner_style_2_m573ef.png" /></a></center></p>


https://www.datacamp.com/community/blog/digest-july-2021

Read our favorite articles from the last month.

AI regulations? Sandboxes seem to be the solution

DataCamp Digest July 2021: AI regulations? Sandboxes seem to be the solution

https://www.datacamp.com/community/blog/the-path-to-building-data-cultures

Sudaman Thoppan Mohanchandralal, Regional Chief Data and Analytics Officer at Allianz Benelux, discusses the importance of building data cultures and his experiences operationalizing data culture transformation programs.

The Path to Building Data Cultures

<p>You&#39;ll probably already know about Apache Spark, the fast, general and open-source engine for big data processing; It has&nbsp;built-in modules for&nbsp;streaming, SQL, machine learning and graph processing. Spark&nbsp;allows you to speed analytic applications up to 100 times faster compared to other technologies on the market today. Interfacing Spark with Python is easy with PySpark: this Spark Python API&nbsp;exposes the Spark programming model to Python.&nbsp;</p>

<p>The <a href="https://www.datacamp.com/community/blog/pyspark-cheat-sheet-python">PySpark Basics cheat sheet</a> already showed you how to work with the most basic building blocks, RDDs.</p>

<p>Now, it&#39;s time to tackle the&nbsp;Spark SQL module, which is meant&nbsp;for structured data processing, and the&nbsp;DataFrame API, which&nbsp;is not only available in Python, but also in Scala, Java, and&nbsp;R. If you want to know more about the differences between RDDs, DataFrames, and DataSets, consider taking a look at <a href="https://www.datacamp.com/community/tutorials/apache-spark-python">Apache Spark in Python: Beginner&#39;s Guide</a>.</p>

<p>Without further ado, here&#39;s the cheat sheet:</p>

<p><a href="http://datacamp-community-prod.s3.amazonaws.com/02213cb4-b391-4516-adcd-57243ced8eed" target="_blank"><img alt="PySpark SQL cheat sheet" src="https://res.cloudinary.com/dyd911kmh/image/upload/f_auto,q_auto:best/v1625837689/PySpark_Cheat_Sheet-_Spark_DataFrames_in_Python_thv8v2.png" style="width: 800px; height: 551px;" /></a></p>

<p>This PySpark SQL cheat sheet covers the basics of working with the&nbsp;Apache Spark DataFrames in Python: from initializing the SparkSession to creating DataFrames, inspecting the data, handling duplicate values, querying, adding, updating or removing columns, grouping, filtering or sorting data. You&#39;ll also see that this cheat sheet also on how to run<b>&nbsp;</b>SQL Queries programmatically, how to save your data to parquet and JSON files, and how to stop your SparkSession.</p>

<p>Make sure&nbsp;to check out our other Python cheat sheets for data science, which cover topics such as&nbsp;<a href="https://www.datacamp.com/community/tutorials/python-data-science-cheat-sheet-basics/">Python basics</a>,&nbsp;<a href="https://www.datacamp.com/community/blog/python-numpy-cheat-sheet/">Numpy</a>,&nbsp;<a href="https://www.datacamp.com/community/blog/python-pandas-cheat-sheet/">Pandas</a>,&nbsp;<a href="https://www.datacamp.com/community/blog/pandas-cheat-sheet-python/">Pandas Data Wrangling</a>&nbsp;and much more!&nbsp;</p>


https://www.datacamp.com/community/blog/pyspark-sql-cheat-sheet

This PySpark SQL cheat sheet is your handy companion to Apache Spark DataFrames in Python and includes code samples.

PySpark Cheat Sheet: Spark DataFrames in Python

<p>Apache Spark is generally known as a fast, general and open-source engine for big data processing, with built-in modules for&nbsp;streaming, SQL, machine learning and graph processing. It allows you to speed analytic applications up to 100 times faster compared to technologies on the market today. You can interface Spark with Python through &quot;PySpark&quot;. This is the Spark Python API&nbsp;exposes the Spark programming model to Python.&nbsp;</p>

<p>Even though working with Spark will remind you in many ways of working with <a href="https://www.datacamp.com/community/tutorials/pandas-tutorial-dataframe-python/">Pandas DataFrames</a>, you&#39;ll also see that it can be tough getting familiar with all the functions that you can use to query, transform,&nbsp;inspect, ...&nbsp;your data. What&#39;s more, if you&#39;ve never worked with any other programming language or if you&#39;re new to the field, it might be hard to distinguish between RDD operations.</p>

<p>Let&#39;s face it, <code>map()</code> and <code>flatMap()</code> are different enough, but it might still come as a challenge to decide which one you really need when you&#39;re faced with them in your analysis. Or what about other functions, like&nbsp;<code>reduce()</code> and <code>reduceByKey()</code>?&nbsp;</p>

<p><a href="http://datacamp-community-prod.s3.amazonaws.com/acfa4325-1d43-4542-8ce4-bea2d287db10" target="_blank"><img alt="PySpark cheat sheet" src="https://res.cloudinary.com/dyd911kmh/image/upload/f_auto,q_auto:best/v1625837623/PySpark_Cheat_Sheet-_Spark_in_Python_owornh.png" style="width: 800px; height: 551px;" /></a></p>

<p>Even though the <a href="http://spark.apache.org/docs/latest/api/python/">documentation</a> is very elaborate, it never hurts to have a cheat sheet by your side, especially when you&#39;re just getting into it.</p>

<p>This PySpark cheat sheet covers the basics, from initializing Spark and loading your data, to retrieving RDD information, sorting, filtering and sampling your data. But that&#39;s not all. You&#39;ll also see that topics such as repartitioning, iterating, merging, saving your data and stopping the SparkContext are included in the cheat sheet.&nbsp;</p>

<p><strong>Note</strong> that the examples in the document take small data sets to illustrate the effect of specific functions on your data. In real life data analysis, you&#39;ll be using Spark to analyze big data.</p>

<p>Are you hungry for more? Don&#39;t miss our other Python cheat sheets for data science that cover topics such as&nbsp;<a href="https://www.datacamp.com/community/tutorials/python-data-science-cheat-sheet-basics/">Python basics</a>,&nbsp;<a href="https://www.datacamp.com/community/blog/python-numpy-cheat-sheet/">Numpy</a>,&nbsp;<a href="https://www.datacamp.com/community/blog/python-pandas-cheat-sheet/">Pandas</a>,&nbsp;<a href="https://www.datacamp.com/community/blog/pandas-cheat-sheet-python/">Pandas Data Wrangling</a>&nbsp;and much more!&nbsp;</p>


https://www.datacamp.com/community/blog/pyspark-cheat-sheet-python

This PySpark cheat sheet with code samples covers the basics like initializing Spark in Python, loading data, sorting, and repartitioning.

PySpark Cheat Sheet: Spark in Python 

<p><center><a target="_blank" href="https://www.datacamp.com/business/demo"><img src="https://res.cloudinary.com/dyd911kmh/image/upload/f_auto,q_auto:best/v1624651228/schedule_a_demo_bfhlx1.png" /></a></p></p>
<p>Driven by desires to reach the elusive goal of becoming data-driven, organizations have been arduously accumulating data, building infrastructure, and attracting highly skilled talents. Considering how data is often touted as the silver bullet to a myriad of problems, these organizations are moving in the right direction.</p>
<p>However, the virtue of having the right tools and talents does not automatically qualify a company as data-driven. The biggest obstacle towards becoming data-driven lies in <a target="_blank" href="https://www.datacamp.com/resources/webinars/building-data-cultures">building a data culture</a>, where data science is seen by everyone as a methodology for solving business problems. Indeed, any organization aspiring to be data-driven needs to first nurture a data culture by fostering organization-wide data fluency, equipping everyone with the skills to perform the data tasks they need to excel at their roles. Such a fundamental shift in mindset presents a sizable challenge in a company’s goal of becoming data-driven. In fact, according to the <a target="_blank" href="https://www.gartner.com/smarterwithgartner/3-top-take-aways-from-the-gartner-chief-data-officer-survey/">2018 Gartner Chief Data Officer Survey</a>, 35% of CDOs believe low data fluency is the biggest challenge towards extracting value from data at scale.</p>
<p>The benefits of becoming data-driven and leveraging data science at scale are numerous and widely documented. Here’s a run down of four areas of data science that organizations can quickly adopt to start extracting value from their data, illustrated with specific use cases.</p>
<h3 id="1-communicating-complex-insights-through-data-visualization-quickly-and-concisely-">1. Communicating complex insights through data visualization, quickly and concisely.</h3>
<p>Interactive data dashboards have become increasingly ubiquitous as they allow stakeholders to easily access up-to-date information.</p>
<p>Dashboards are incredibly versatile. They can be a telescope for a company’s long-term north star metrics or a microscope that intently focuses on its short-term operational details. Either way, they allow organizations to scale their data driven decision making.
Today, many no-code business intelligence tools like <a target="_blank" href="https://www.datacamp.com/tracks/tableau-fundamentals">Tableau</a> and <a target="_blank" href="https://www.datacamp.com/courses/introduction-to-power-bi">PowerBI</a> boast intuitive drag-and-drop interfaces for building powerful dashboards. On the other hand, open-source packages built on popular programming languages, like <a target="_blank" href="https://www.datacamp.com/courses/introduction-to-data-visualization-with-plotly-in-python">Python’s plotly</a> and <a target="_blank" href="https://www.datacamp.com/tracks/shiny-fundamentals-with-r">R’s Shiny</a>, provide organizations a low barrier to entry when building highly customizable interactive visualizations.</p>
<h3 id="2-make-better-decisions-by-analyzing-historical-data">2. Make better decisions by analyzing historical data</h3>
<p>A data-driven company describes, summarizes, and understands its historical performance and leverages it to guide decisions that enable it to double-down on what works and course-correct what doesn’t.. A survey by Deloitte found that about half of the companies assert that the most significant benefit to analytics is to <a target="_blank" href="https://www2.deloitte.com/content/dam/Deloitte/global/Documents/Deloitte-Analytics/dttl-analytics-analytics-advantage-report-061913.pdf">enable better data-driven decision-making</a>. Also, 62% of the companies said data is important to drive business strategy. Clearly, data is becoming an increasingly irreplaceable instrument in the corporate toolbox.</p>
<p><a target="_blank" href="https://www.datacamp.com/courses/data-driven-decision-making-for-business">Data-driven decision making</a> can maintain and enhance an organization’s competitive advantage over its competitors. For example, <a target="_blank" href="https://www.datacamp.com/courses/market-basket-analysis-in-python">market basket analysis</a> can help physical retailers better optimize  their physical spaces and create better shopping experiences for their customers. Marketers can analyze historical email click through rates to evaluate the success of email marketing campaigns. Financial analysts can apply <a target="_blank" href="https://www.datacamp.com/courses/introduction-to-time-series-analysis">time-series analysis</a> to their historical data to optimize budget planning.  </p>
<p>Such analysis of historical data can be done efficiently on spreadsheet tools like Microsoft Excel and Google Sheets when the size of the data is small. As the data grows, analytics can be done most effectively using open-source programming languages like <a target="_blank" href="https://www.datacamp.com/community/blog/how-to-learn-r">R</a> and <a target="_blank" href="https://www.datacamp.com/community/blog/how-to-learn-python">Python</a>.</p>
<h3 id="3-supplement-existing-subject-matter-expertise-with-statistical-thinking">3. Supplement Existing Subject Matter Expertise with Statistical Thinking</h3>
<p>Statistics can be used to succinctly summarize large amounts of data. At the very foundation of statistics are summary statistics that represent averages (like mean, mode and median) and deviations from average (like standard deviation and variance). These summary statistics are the building blocks of tools like the box plot and metrics like correlation coefficient (a measure of the strength of the relationship between two variables). For example, retailers can calculate the correlation between the purchase of products to optimize cross-selling campaigns. Market researchers can leverage conjoint analysis to determine the ideal mix of customer preferences before launching a new product feature or changing the pricing of a product.</p>
<p>Controlled experiments can help companies isolate the effect of a change, establish a clear cause-and-effect relationship between two metrics, and make better decisions. With foundations rooted in statistics, controlled experiments provide a relatively simple yet rigorous method of verifying or rejecting the gut instincts of subject matter experts. A simple illustration of a controlled experiment includes <a href="https://www.datacamp.com/courses/ab-testing-in-r#!]">A/B testing</a> email subject lines and choosing the one with the highest click through rate. At AirBnB, only when an experiment conclusively shows that <a target="_blank" href="https://medium.com/airbnb-engineering/experiments-at-airbnb-e2db3abf39e7">a change in the maximum price filter causes conversion rate</a> to increase was the change implemented to all users. At Shutterstock, the idea to remove watermarks on its search page images was rejected when <a target="_blank" href="https://hbr.org/2014/02/ab-testing-and-the-benefits-of-an-experimentation-culture">experiments provided strong evidence</a> that such a change decreased conversion rate.</p>
<h3 id="4-make-predictions-and-discover-new-knowledge-with-machine-learning">4. Make predictions and discover new knowledge with machine learning</h3>
<p>Companies can use past data to make predictions about the future with <a target="_blank" href="https://www.datacamp.com/tracks/machine-learning-scientist-with-python">machine learning</a>. For example, by leveraging supervised learning, a technique that trains algorithms to classify data or predict outcomes with labeled data, any organization can predict which customers are the most likely to <a target="_blank" href="https://www.datacamp.com/courses/marketing-analytics-predicting-customer-churn-in-python">churn</a> based on their historical activity. A trading firm can use historical prices of a stock to predict future prices in financial modelling.</p>
<p>In contrast with supervised learning, unsupervised learning is a technique that can be used by a company with little knowledge about a raw data set to explore its data—even if it does not know what it is looking for.Clustering is an example of an unsupervised learning algorithm that sorts data into categories that are not predetermined. For instance, a telecom company that has valuable data about its customers’ data and has yet to find a way to segment its customers can use <a target="_blank" href="https://www.datacamp.com/courses/customer-segmentation-in-python">unsupervised learning</a> to identify customer clusters with distinctive characteristics.</p>
<p>Unsupervised learning is also <a target="_blank" href="https://www.forbes.com/sites/davidteich/2020/09/03/management-ai-matching-ai-models-to-business-needs-unsupervised-learning-customer-segmentation-and-association/?sh=5e00b5f345f4">superb at finding associations</a>. In e-commerce, product recommendations in the form of ‘people who like product X also like product Y’ are often a result of unsupervised learning techniques that uncover such hidden linkages. Such product recommendations can form the backbone of an effective cross-selling strategy.</p>
<h3 id="building-a-data-driven-organization">Building a Data-Driven Organization</h3>
<p>Evidently, making the steps towards becoming data-driven unlocks many potential benefits for an organization. There are many low-hanging fruit data use-cases that functional teams can start applying together, requisite they have the necessary skills to get started. That is why <a target="_blank" href="https://www.datacamp.com/groups/business">DataCamp for Business</a> provides an interactive learning platform for companies that need to upskill and reskill their people on data skills. With topics ranging from data literacy, and data science to data engineering, and machine learning, over 1,600 companies trust DataCamp for Business to upskill their talent.</p>
<p><center><a target="_blank" href="https://www.datacamp.com/business/demo"><img src="https://res.cloudinary.com/dyd911kmh/image/upload/f_auto,q_auto:best/v1624651228/schedule_a_demo_bfhlx1.png" /></a></p></p>


https://www.datacamp.com/community/blog/four-ways-your-team-can-start-leveraging-data-science

Becoming a data-driven organization can significantly help you make more effective resource allocation decisions, but this requires nurturing and fostering a company-wide data culture. Here’s a rundown of various data science practices that you can quickly adapt to start extracting value for your organization

Here’s a rundown of various data science practices that you can quickly adapt to start extracting value for your organization

community

datacamp

Official Blog