Data Science Using R Training

1 Introduction to Data Science

Part 1

What is Data science and it’s 5 disruptions

Data science v traditional methods

Difference in architecture, reference architecture

Demystifying machine learning

Segmentation technique using R

Kmeans, ggplot, ScatterPlot commands

Part 2

Basic R commands

Assigning values to objects

Creating vectors, matrices

Importing data into R, packages to R

RStudio basic options

Boxplot, pie, bar chart commands

2 Signaling concepts

Signals – Key Concepts

Analyzing a signal pattern

Signal extraction methodology

Simplistic nine step process

Commands in R – setwd, Dim, Table, Str

Internalize meta-model using commands

Assignment 1: Signaling concepts

3 Uni-Analysis: Commands, Functions in R / Assignment 1

Part 1

Uni-variate Analysis

Fleet data analysis using Uni-variate concepts

Uni-variate outputs using R

Using Summary, Table, GGPlot commands

Assignment 2: Use summary, table, ggplot commands

Part 2

Concepts of Sample, Population

Hypothesis testing: Null and alternate

Significance levels/P value

Probabilities calculation

pnorm, qnorm, dnorm functions

abline, Rnorm commands

4 Bi-Analysis: Commands, Concepts in R

Visual construct using box, scatter plots, Geo-spatial, heat maps

Heats maps example using fillets, brewing industry

Spider charts

Domestic loan analysis

Core concepts in advanced visualization: visualization consumers

Creating dashboards

Visualization commands in R: Plot, Boxplot, Scatter. smooth, pairs, sp commands

5 Visualization with R

Visual construct using box, scatter plots, Geo-spatial, heat maps

Heats maps example using fillets, brewing industry

Spider charts

Domestic loan analysis

Core concepts in advanced visualization: visualization consumers

Creating dashboards

Visualization commands in R: Plot, Boxplot, Scatter. smooth, pairs, sp commands

6 Advanced visualization with R

Business story telling using R

Small multiples, bubble charts commands in R

Library command to display libraries

Union command to merge databases

Unique command to remove duplicate information

Intersect command to find common information in two datasets

7 Case study: Exploratory Data Analysis (EDS) with R

Scenario 1: Survival Analysis

Scenario 2: Attrition Analysis

Scenario 3: Valuable Vulnerable

Scenario 4: Day to Repeat Purchase

Scenario 5: Identifying Patterns

Scenario 6: Segmenting Watch Companies

Scenario 7: Customer Lifetime Value

8 Machine Learning in Action

Support Vector Machines (SVM), Decision Trees, Random Forest algorithms

A/B Testing

Collaborative Filtering

Fixed Size, Threshold based Neighborhood

Graphs

Applying algorithms to structured, unstructured data

9 Regression

Part 1

5 powerful unanswered questions by regression which remain unknown

Regression Across Sectors

Scenario 1: Cost of Insurance

Scenario 2: Model Building for Property Design

Scenario 3: Estimating Patients Stay at Hospital

Scenario 4: Estimate Defect Density

• Population, Sample Regression Models
• Commands in R
• Correlation # Causation

Part 2

        Linear regression and dependent variables

        Lm command

        Summary of models

        Attribute extraction, assumptions made while fitting a linear model

        Diagnostic plots in R

10 Dimensionality Reduction Techniques

Feature Engineering – Key Point

Feature Selection—Definition

Feature Selection—Optimality

Ranking Criteria—Correlation

Feature Subset Selection