Data Science Training Course

Description: Data science is a multi-disciplinary field that uses scientific methods, processes, algorithms and systems to extract knowledge and insights from structured and unstructured data. In this first module we will introduce to the field of Data Science and how it relates to other fields of data like Artificial Intelligence, Machine Learning and Deep Learning.

• Introduction to Data Science
• High level view of Data Science, Artificial Intelligence & Machine Learning
• Subtle differences between Data Science, Machine Learning & Artificial Intelligence
• Approaches to Machine Learning
• Terms & Terminologies of Data Science
• Understanding an end to end Data Science Pipeline, Implementation cycle

Mathematics is very important in the field of data science as concepts within mathematics aid in identifying patterns and assist in creating algorithms. The understanding of various notions of Statistics and Probability Theory are key for the implementation of such algorithms in data science.

• Linear Algebra
• Matrices, Matrix Operations
• Eigen Values, Eigen Vectors
• Scalar, Vector and Tensors
• Prior and Posterior Probability
• Conditional Probability
• Calculus
• Differentiation, Gradient and Cost Functions
• Graph Theory

This module focuses on understanding statistical concepts required for Data Science, Machine Learning and Deep Learning. In this module, you will be introduced to the estimation of various statistical measures of a data set, simulating random distributions, performing hypothesis testing, and building statistical models.

Descriptive Statistics

• Types of Data (Discrete vs Continuous)
• Types of Data (Nominal, Ordinal)
• Measures of Central Tendency (Mean, Median, Mode)
• Measures of Dispersion (Variance, Standard Deviation)
• Range, Quartiles, Inter Quartile Ranges
• Measures of Shape (Skewness and Kurtosis)
• Tests for Association (Correlation and Regression)
• Random Variables
• Probability Distributions
• Standard Normal Distribution
• Probability Distribution Function
• Probability Mass Function
• Cumulative Distribution Function
• Inferential Statistics
• Statistical sampling & Inference
• Hypothesis Testing
• Null and Alternate Hypothesis
• Margin of Error
• Type I and Type II errors
• One Sided Hypothesis Test, Two-Sided Hypothesis Test
• Tests of Inference: Chi-Square, T-test, Analysis of Variance
• t-value and p-value
• Confidence Intervals

Python for Data Science

• Numpy
• Pandas
• Matplotlib & Seaborn
• Jupyter Notebook

Numpy

NumPy is a Python library that works with arrays when performing scientific computing with Python. Explore how to initialize and load data into arrays and learn about basic array manipulation operations using NumPy.

• Loading data with Numpy
• Comparing Numpy with Traditional Lists
• Numpy Data Types
• Indexing and Slicing
• Copies and Views
• Numerical Operations with Numpy
• Matrix Operations on Numpy Arrays
• Aggregations functions
• Shape Manipulations
• Broadcasting
• Statistical operations using Numpy
• Resize, Reshape, Ravel
• Image Processing with Numpy

Pandas

Pandas is a Python library that provides utilities to deal with structured data stored in the form of rows and columns. Discover how to work with series and tabular data, including initialization, population, and manipulation of Pandas Series and DataFrames.

• Basics of Pandas
• Loading data with Pandas
• Series
• Operations on Series
• DataFrames and Operations of DataFrames
• Selection and Slicing of DataFrames
• Descriptive statistics with Pandas
• Map, Apply, Iterations on Pandas DataFrame
• Working with text data
• Multi Index in Pandas
• GroupBy Functions
• Merging, Joining and Concatenating DataFrames
• Visualization using Pandas

Data Visualization using Matplotlib

Matplotlib is a plotting library for the Python programming language and its numerical mathematics extension NumPy. It provides an object-oriented API for embedding plots into applications using general-purpose GUI toolkits like Tkinter, wxPython, Qt, or GTK+

• Anatomy of Matplotlib figure
• Plotting Line plots with labels and colors
• Adding markers to line plots
• Histogram plots
• Scatter plots
• Size, Color and Shape selection in Scatter plots.
• Applying Legend to Scatter plotsDisplaying multiple plots using subplots
• Boxplots, scatter_matrix and Pair plots

Data Visualization using Seaborn

• Seaborn is a data visualization library that provides a high-level interface for drawing graphs. These graphs are able to convey a lot of information, while also being visually appealing.
• Basic Plotting using Seaborn
• Violin Plots
• Box Plots
• Cat Plots
• Facet Grid
• Swarm Plot
• Pair Plot
• Bar Plot
• LM Plot
• Variations in LM plot using hue, markers, row and col

Exploratory Data Analysis helps in identifying the patterns in the data by using basic statistical methods as well as using visualization tools to displays graphs and charts. With EDA we can assess the distribution of the data and conclude various models to be used.

Pipeline ideas

• Exploratory Data Analysis
• Feature Creation
• Evaluation Measures

Data Analytics Cycle ideas

• Data Acquisition
• Data Preparation
• Data cleaning
• Data Visualization
• Plotting
• Model Planning & Model Building

Data preparation

• Selection and Removal of Columns
• Transform
• Rescale
• Standardize
• Normalize
• Binarize
• One hot Encoding
• Imputing
• Train, Test Splitting

In machine learning, computers apply statistical learning techniques to automatically identify patterns in data. This module on Machine Learning is a deep dive to Supervised, Unsupervised learning and Gaussian / Naive-Bayes methods. Also you will be exposed to different classification, clustering and regression methods.

• Introduction to Machine Learning
• Applications of Machine Learning
• Supervised Machine Learning
• Classification
• Regression
• Unsupervised Machine Learning
• Reinforcement Learning
• Latest advances in Machine Learning
• Model Representation
• Model Evaluation
• Hyper Parameter tuning of Machine Learning Models.
• Evaluation of ML Models.
• Estimating and Prediction of Machine Learning Models
• Deployment strategy of ML Models.

Supervised learning is one of the most popular techniques in machine learning. In this module, you will learn about more complicated supervised learning models and how to use them to solve problems.

Classification methods & respective evaluation

• K Nearest Neighbors
• Decision Trees
• Naive Bayes
• Stochastic Gradient Descent
• SVM –
• Linear
• Non linear
• Radial Basis Function
• Random Forest
• Gradient Boosting Machines
• XGboost
• Logistic regression

Ensemble methods

• Combining models
• Bagging
• Boosting
• Voting
• Choosing best classification method

Model Tuning

• Train Test Splitting
• K-fold cross validation
• Variance bias tradeoff
• L1 and L2 norm
• Overfit, underfit along with learning curves variance bias sensibility using graphs
• Hyper Parameter Tuning using Grid Search CV

Respective Performance measures

• Different Errors (MAE, MSE, RMSE)
• Accuracy, Confusion Matrix, Precision, Recall

Regression is a type of predictive modelling technique which is heavily used to derive the relationship between variables (the dependent and independent variables). This technique finds its usage mostly in forecasting, time series modelling and finding the causal effect relationship between the variables. The module discusses in detail about regression and types of regression and its usage & applicability

Regression

• Linear Regression
• Variants of Regression
• Lasso
• Ridge
• Multi Linear Regression
• Logistic Regression (effectively, classification only)
• Regression Model Improvement
• Polynomial Regression
• Random Forest Regression
• Support Vector Regression
• Respective Performance measures
• Different Errors (MAE, MSE, RMSE)
• Mean Absolute Error
• Mean Square Error
• Root Mean Square Error

Unsupervised learning can provide powerful insights on data without the need to annotate examples. In this module, you will learn several different techniques in unsupervised machine learning.

Clustering

• K means
• Hierarchical Clustering
• DBSCAN

Association Rule Mining

• Association Rule Mining.
• Market Basket Analysis using Apriori Algorithm
• Dimensionality reduction using Principal Component analysis (PCA)

Natural language is essential to human communication, which makes the ability to process it an important one for computers. In this module, you will be introduced to natural language processing and some of the basic tasks.

• Text Analytics
• Stemming, Lemmatization and Stop word removal.
• POS tagging and Named Entity Recognition
• Bigrams, Ngrams and colocations
• Term Document Matrix
• Count Vectorizer
• Term Frequency and TF-IDF

Advanced Analytics covers various areas like Time series Analysis, ARIMA models, Recommender systems etc.

• Time series
• Time series Analysis.
• ARIMA example
• Recommender Systems
• Content Based Recommendation
• Collaborative Filtering

Reinforcement learning is an area of Machine Learning which takes suitable action to maximize reward in a particular situation. It is employed by various software and machines to find the best possible behavior or path it should take in a specific situation.

Basic concepts of Reinforcement Learning

• Action
• Reward
• Penalty Mechanism
• Feedback loop
• Deep Q Learning

Artificial intelligence (AI) is the ability of a computer program or a machine to think and learn. It is also a field of study which tries to make computers "smart"

Artificial Neural Networks

• Neural Networks & terminologies
• Non linearity problem, illustration
• Perceptron learning
• Feed Forward Network and Back propagation
• Gradient Descent

Mathematics of Artificial Neural Networks

• Gradients
• Partial derivatives
• Linear algebra
• Li
• LD
• Eigen vectors
• Projections
• Vector quantization

Overview of tools used in Neural Networks

Tensor Flow

Keras

Deep learning is part of a broader family of machine learning methods based on the layers used in artificial neural networks. In this module, you’ll deep dive in the concepts of Neural Networks, Convolutional Neural Networks, Recurrent Neural Networks, Auto Encoders and many more.

Deep Learning

• Tensorflow & keras installation
• More elaborate discussion on cost function
• Measuring accuracy of hypothesis function
• Role of gradient function in minimizing cost function
• Explicit discussion of Bayes models
• Hidden Markov Models (HMM)
• Optimization basics
• Sales Prediction of a Gaming company using Neural Networks
• Build an Image similarity engine.

Deep Learning with Convolutional Neural Nets

• Architecture of CNN
• Types of layers in CNN
• Different Filters and Kernels
• Building an Image classifier with and without CNN

Recurrent neural nets

• Fundamental notions & ideas
• Recurrent neurons
• Handling variable length sequences
• Training a sequence classifier
• Training to predict Time series

Cloud computing is massively growing in importance in the IT sector as more and more companies are eschewing traditional IT and moving applications and business processes to the cloud. This section covers detailed information about how to deploy Data Science models on Cloud environments.

Topics

• Introduction to Cloud Computing
• Amazon Web Services Preliminaries - S3, EC2, RDS
• Big data processing on AWS using Elastic Map Reduce (EMR)
• Machine Learning using Amazon Sage Maker
• Deep Learning on AWS Cloud
• Natural Language processing using AWS Lex
• Analytics services on AWS Cloud
• Data Warehousing on AWS Cloud