Arani Consulting

Email:info@araniconsulting.com

Unit 1: Introduction to Data Science with R

1. What is Data Science, the significance of Data Science in today’s digitally-driven world, applications of Data Science, lifecycle of Data Science, components of the Data Science lifecycle, introduction to big data and Hadoop, introduction to Machine Learning and Deep Learning, introduction to R programming and R Studio.

Unit 2: Data Exploration

1. Introduction to data exploration, importing and exporting data to/from external sources, what is data exploratory analysis, data importing, dataframes, working with dataframes, accessing individual elements, vectors and factors, operators, in-built functions, conditional, looping statements and user-defined functions, matrix, list and array.

Unit 3: Data Manipulation

1. Need for Data Manipulation, Introduction to dplyr package, Selecting one or more columns with select() function, Filtering out records on the basis of a condition with filter() function, Adding new columns with the mutate() function, Sampling & Counting with sample_n(), sample_frac() & count() functions, Getting summarized results with the summarise() function, Combining different functions with the pipe operator, Implementing sql like operations with sqldf.

Unit 4: Data Visualization

1. Introduction to visualization, Different types of graphs, Introduction to grammar of graphics & ggplot2 package, Understanding categorical distribution with geom_bar() function, understanding numerical distribution with geom_hist() function, building frequency polygons with geom_freqpoly(), making a scatter-plot with geom_pont() function, multivariate analysis with geom_boxplot, univariate Analysis with Bar-plot, histogram and Density Plot, multivariate distribution, Bar-plots for categorical variables using geom_bar(), adding themes with the theme() layer, visualization with plotly package & building web applications with shinyR, frequency-plots with geom_freqpoly(), multivariate distribution with scatter-plots and smooth lines, continuous vs categorical with box-plots, subgrouping the plots, working with co-ordinates and themes to make the graphs more presentable, Intro to plotly & various plots, visualization with ggvis package, geographic visualization with ggmap(), building web applications with shinyR.

Unit 5: Introduction to Statistics

1. Why do we need Statistics?, Categories of Statistics, Statistical Terminologies,Types of Data, Measures of Central Tendency, Measures of Spread, Correlation & Covariance,Standardization & Normalization,Probability & Types of Probability, Hypothesis Testing, Chi-Square testing, ANOVA, normal distribution, binary distribution.

Unit 6: Machine Learning

1. Introduction to Machine Learning, introduction to Linear Regression, predictive modeling with Linear Regression, simple Linear and multiple Linear Regression, concepts and formulas, assumptions and residual diagnostics in Linear Regression, building simple linear model, predicting results and finding p-value, introduction to logistic regression, comparing linear regression and logistics regression, bivariate & multi-variate logistic regression, confusion matrix & accuracy of model, threshold evaluation with ROCR, Linear Regression concepts and detailed formulas, various assumptions of Linear Regression,residuals, qqnorm(), qqline(), understanding the fit of the model, building simple linear model, predicting results and finding p-value, understanding the summary results with Null Hypothesis, p-value & F-statistic, building linear models with multiple independent variables.

Unit 7: Logistic Regression

1. Introduction to Logistic Regression, Logistic Regression Concepts, Linear vs Logistic regression, math behind Logistic Regression, detailed formulas, logit function and odds, Bi-variate logistic Regression, Poisson Regression, building simple “binomial” model and predicting result, confusion matrix and Accuracy, true positive rate, false positive rate, and confusion matrix for evaluating built model, threshold evaluation with ROCR, finding the right threshold by building the ROC plot, cross validation & multivariate logistic regression, building logistic models with multiple independent variables, real-life applications of Logistic Regression.

Unit 8: Decision Trees & Random Forest

1. What is classification and different classification techniques, introduction to Decision Tree, algorithm for decision tree induction, building a decision tree in R, creating a perfect Decision Tree, Confusion Matrix, Regression trees vs Classification trees, introduction to ensemble of trees and bagging, Random Forest concept, implementing Random Forest in R, what is Naive Bayes, Computing Probabilities, Impurity Function – Entropy, understand the concept of information gain for right split of node, Impurity Function – Information gain, understand the concept of Gini index for right split of node, Impurity Function – Gini index, understand the concept of Entropy for right split of node, overfitting & pruning, pre-pruning, post-pruning, cost-complexity pruning, pruning decision tree and predicting values, find the right no of trees and evaluate performance metrics.

Unit 9: Unsupervised learning

1. What is Clustering & it’s Use Cases, what is K-means Clustering, what is Canopy Clustering, what is Hierarchical Clustering, introduction to Unsupervised Learning, feature extraction & clustering algorithms, k-means clustering algorithm, Theoretical aspects of k-means, and k-means process flow, K-means in R, implementing K-means on the data-set and finding the right no. of clusters using Scree-plot, hierarchical clustering & Dendogram, understand Hierarchical clustering, implement it in R and have a look at Dendograms, Principal Component Analysis, explanation of Principal Component Analysis in detail, PCA in R, implementing PCA in R.

Unit 10: Association Rule Mining & Recommendation Engine

1. Introduction to association rule Mining & Market Basket Analysis, measures of Association Rule Mining: Support, Confidence, Lift, Apriori algorithm & implementing it in R, Introduction to Recommendation Engine, user-based collaborative filtering & Item-Based Collaborative Filtering, implementing Recommendation Engine in R, user-Based and item-Based, Recommendation Use-cases.

Unit 11: Introduction to Artificial Intelligence

1. Introducing Artificial Intelligence and Deep Learning, what is an Artificial Neural Network, TensorFlow – computational framework for building AI models, fundamentals of building ANN using TensorFlow, working with TensorFlow in R.

Unit 12: Time Series Analysis

1. What is Time Series, techniques and applications, components of Time Series, moving average, smoothing techniques, exponential smoothing, univariate time series models, multivariate time series analysis, Arima model, Time Series in R, sentiment analysis in R (Twitter sentiment analysis), text analysis.

Unit 13: Support Vector Machine – (SVM)

1. Introduction to Support Vector Machine (SVM), Data classification using SVM, SVM Algorithms using Separable and Inseparable cases, Linear SVM for identifying margin hyperplane.

Unit 14: Naïve Bayes

1. What is Bayes theorem, What is Naïve Bayes Classifier, Classification Workflow, How Naive Bayes classifier works, Classifier building in Scikit-learn, building a probabilistic classification model using Naïve Bayes, Zero Probability Problem.

Unit 15: Text Mining

1. Introduction to concepts of Text Mining, Text Mining use cases, understanding and manipulating text with ‘tm’ & ‘stringR’, Text Mining Algorithms, Quantification of Text, Term Frequency-Inverse Document Frequency (TF-IDF), After TF-IDF.

Case Study

Case Study 1: The Market Basket Analysis (MBA)

1. This case study is associated with the modeling technique of Market Basket Analysis where you will learn about loading of data, various techniques for plotting the items and running the algorithms. It includes finding out what are the items that go hand in hand and hence can be clubbed together. This is used for various real world scenarios like a supermarket shopping cart and so on.

Case Study 2: Logistic Regression

1. In this case study you will get a detailed understanding of the advertisement spends of a company that will help to drive more sales. You will deploy logistic regression to forecast future trends, detect patterns, uncover insights and more all through the power of R programming. Due to this the future advertisement spends can be decided and optimized for higher revenues.

Case Study 3: Multiple Regression

1. You will understand how to compare the miles per gallon (MPG) of a car based on the various parameters. You will deploy multiple regression and note down the MPG for car make, model, speed, load conditions, etc. It includes the model building, model diagnostics, checking the ROC curve, among other things.

Case Study 4: Receiver Operating Characteristic (ROC)

1. You will work with various data sets in R, deploy data exploration methodologies, build scalable models, predict the outcome with the highest precision, diagnose the model that you have created with various real-world data, check the ROC curve and more.

Data Science Projects

Project 1 : Augmenting retail sales with Data Science

• Industry : Retail
• Problem Statement : How to deploy the various rules and algorithms of Data Science for analyzing stationary store purchase data.
• Topics : In this project you will deploy the various tools of Data Science like association rule, Apriori algorithm in R, support, lift and confidence of association rule. You will analyze the purchase data of the stationary outlet for three days and understand the customer buying patterns across products.
• Highlights:
  1. Association rules for transaction data
  2. Association mining with Apriori algorithm
  3. Generating rules and identifying patterns

Project 2 : Analyzing pre-paid model of stock broking

• Industry : Finance
• Problem Statement : Finding out the deciding factor for people to opt for the pre-paid model of stock broking.
• Topics : In this Data Science project you will learn about the various variables that are highly correlated in pre-paid brokerage model, analysis of various market opportunities, developing targeted promotion plans for various products sold under various categories. You will also do competitor analysis, the advantages and disadvantages of pre-paid model.
• Highlights:
  1. Deploying the rules of statistical analysis
  2. Implementing data visualization
  3. Linear regression for predictive modeling.

Project 3 : Cold Start Problem in Data Science

• Industry : Ecommerce
• Problem Statement : How to build a recommender system without the historical data available
• Topics : This project involves understanding of the cold start problem associated with the recommender systems. You will gain hands-on experience in information filtering, working on systems with zero historical data to refer to, as in the case of launching a new product. You will gain proficiency in working with personalized applications like movies, books, songs, news and such other recommendations. This project includes the various ways of working with algorithms and deploying other data science techniques.
• Highlights:
  1. Algorithms for Recommender
  2. Ways of Recommendation
  3. Types of Recommendation -Collaborative Filtering Based Recommendation, Content-Based Recommendation
  4. Complete mastery in working with the Cold Start Problem.

Project 4 : Recommendation for Movie, Summary

• Industry : Ecommerce
• Topics : This is real world project that gives you hands-on experience in working with a movie recommender system. Depending on what movies are liked by a particular user, you will be in a position to provide data-driven recommendations. This project involves understanding recommender systems, information filtering, predicting ‘rating’, learning about user ‘preference’ and so on. You will exclusively work on data related to user details, movie details and others. The main components of the project include the following:
  1. Recommendation for movie
  2. Two Types of Predictions – Rating Prediction, Item Prediction
  3. Important Approaches: Memory Based and Model-Based
  4. Knowing User Based Methods in K-Nearest Neighbor
  5. Understanding Item Based Method
  6. Matrix Factorization
  7. Decomposition of Singular Value
  8. Data Science Project discussion
  9. Collaboration Filtering
  10. Business Variables Overview

Project 5 : Prediction on Pokemon dataset

• Industry : Gaming
• Problem Statement : For the purpose of this case study, you are a Pokemon trainer who is on his way to catch all the 800 Pokemons
• Topics : This real-world project will give you a hands-on experience on the data science life cycle. You’ll understand the structure of the ‘Pokemon’ dataset & use machine learning algorithms to make some predictions. You will use the dplyr package to filter out specific Pokemons and use decision trees to find if the Pokemon is legendary or not.
• Highlights:
  1. dplyr package to filter Pokemons
  2. Decision Tree algorithm
  3. Linear regression algorithm.

Project 6 : Book Recommender System

• Industry : E-commerce
• Problem Statement : Building a book recommender system for readers with similar interests
• Topics : This real-world project will give you a hands-on experience in working with a book recommender system. Depending on what books are read by a particular user, you will be in a position to provide data-driven recommendations. You will understand the structure of the data and visualize it to find interesting patterns.
• Highlights:
  1. Data analysis & visualization
  2. Recommender Lab
  3. User Based Collaborative Filtering Model.

Project 7 : Capstone

• Industry : Analytics
• Problem Statement : Predicting if the customer will churn or not.
• Topics : An end-to-end capstone project comprising:
  1. Manipulating and envisioning the data for insights.
  2. Implementing the linear regression model to predict continuous values.
  3. Implementing classification models – decision tree, logistic regression, and random forest on “customer churn”.
• Highlights:
  1. An end-to-end capstone project covering all the modules. You’ll start off by manipulating and visualizing the data to get interesting insights. Then you’d have to implement the linear regression model to predict continuous values. Following which you’ll implement these classification models – logistic regression, decision tree & random forest on the “customer churn” data frame to find if the customer will churn or not.