Intro to AI with Machine Learning, Deep Learning and More

Natural Language Processing

• Perform natural language processing (NLP) tasks, which are fundamental to many of the forthcoming data science case studies
• Use the textblob, nltk, textatistic and spacy NLP libraries and their pretrained models to perform various NLP tasks
• Tokenize text
• Parts-of-speech tagging
• Sentiment analysis
• Language detection and translation
• Stemming and lemmatization
• Spell checking and correction capabilities
• Getting word definitions, synonyms and antonyms
• Removing stop words from text
• Creating word clouds with the wordcloud library
• Determining text readability with textatistic
• Using the spacy library for named entity recognition and similarity detection

Data Mining Twitter

• Twitter’s impact on businesses, brands, reputation, sentiment analysis, predictions and more
• Register for a free Twitter developer account and get credentials to use their free web services
• Use the tweepy Twitter API client to data mine Twitter
• Search past tweets
• Get a live stream of tweets
• Access tweet metadata
• Use natural language processing to prepare tweets for analysis
• Perform sentiment analysis on tweets
• Spot trending topics
• Map tweets using the geopy library to geocode locations into latitudes and longitudes, and the folium library with OpenStreetMap map tiles

IBM Watson and Cognitive Computing

• Learn Watson’s range of services via their free Lite tier
• Try demos of Watson services
• Register for a free IBM Cloud account and get credentials to use various services
• Install the Watson Developer Cloud Python SDK to interact with Watson services
• Develop a traveler’s companion language translator app by using Python to weave together a mashup of the Watson Speech to Text, Language Translator and Text to Speech services

Machine Learning with scikit-learn

• Use the scikit-learn (sklearn) library with popular datasets to perform machine learning studies
• Visualize and explore data with the seaborn and matplotlib libraries
• Perform supervised machine learning with k-nearest neighbors classification and linear regression
• Perform multi-classification with the Digits dataset
• Divide a dataset into training, test and validation sets
• Tune model hyperparameters with k-fold cross-validation
• Measure model performance
• Display a confusion matrix showing classification prediction hits and misses
• Perform multiple linear regression with the California Housing dataset
• Perform dimensionality reduction with PCA and TSNE on the Iris and Digits datasets to prepare them for two-dimensional visualizations
• Perform unsupervised machine learning with k-means clustering and the Iris dataset

Deep Learning

• Neural networks and how they enable deep learning
• Create Keras neural networks in the context of TensorFlow
• Keras layers, activation functions, loss functions and optimizers
• Deep Learning for Computer Vision: Convolutional neural network (CNN) trained on the MNIST dataset to recognize handwritten digits
• Deep Learning for Sentiment Analysis—Recurrent neural network (RNN) trained on the IMDb (Internet Movie Database) dataset to perform binary classification of positive and negative movie reviews
• Visualize a deep-learning network’s training progress with TensorBoard

Big Data: Relational Databases

• Structured Query Language (SQL), SQLite with the sqlite module

Big Data: NoSQL

• A brief tour of NoSQL and NewSQL big-data databases
• Case Study: A MongoDB JSON document database
• Creating the MongoDB Atlas cluster
• Streaming tweets into MongoDB

Big Data: Apache Spark and the pyspark Library

• Docker and the Jupyter Docker Stacks
• Word count with Spark
• Spark Streaming: Counting Twitter hashtags using the pyspark-notebook Docker stack

Big Data: Internet of Things (IoT) and Dashboards

• Communicating among apps and IoT devices via publish/subscribe
• Visualizing streaming data with freeboard.io and seaborn
• IoT: Simulating an Internet-connected thermostat in Python with dweepy and dweet.io
• Simulated streaming stock prices with pubNub