Data Scientist Introduction bref overview of Concepts

Data Science Introduction

1. Data Scientist April 2018

2. Agenda Data Science and its Application Stages-Data science, Project roles C lassification , decision tree , random forest Demo using R Technology

3. Data Science Introduction-Concepts D ata science is managing the process that can transform hypotheses and data into actionable predictions .

4. Data Science Introduction-Applications Amazon’s product recommendation systems LinkedIn’s contact recommendation system Retail Business – Buying patterns , segment Twitter’s trending topics Google’s advertisement valuation systems Walmart’s consumer demand projection systems

5. Data Science Domains

6. Data Science Introduction-Project Roles

7. Data Science Introduction- Processes in Data Science Project

8. Data Science – Modelling Methods

9. Data Science – Modelling Method Classification and Regression Trees Example :- Finding bad l oan applications Input variables :- Loan amount, duration, a ge, salary , any other loan , address, I ncome , education , background data , location etc 1000 applications exist out of which 300 have been defaulted Decision Tree for identifying Potential defaulters

10. Classification and Regression Trees Duration>50 Amount>4 million Amount> 1mil Amount<5 mil Bad (0.68) Duration>120 Good (0.75) Good (0.56) Bad (0.25) Good (0.61) Bad (0.88)

11. Data Science – Modelling Method K – nearest Neighbors( Knn ) Example : Male , Female distribution Hair Length ( c ms ) 60 40 20 0/ 140 150 160 170 180 190 200 Height ( cms )

12. Data Science – Modelling Method K – nearest Neighbors( Knn ) Example : Male , Female distribution Hair Length ( c ms ) 60 40 20 0/ 140 150 160 170 180 190 200 Height ( cms )

13. Data Science – Modelling Method Random Forest (RF) Tree 1 Tree 3 Tree 2

14. Data Science – Modelling Method Random Forest (RF) Input All Trees Prediction Tree1: Tree2: Tree3: Random Forest Predicts:

15. Data Science – Modelling Method Random Forest (RF) A pplication where random forest algorithm is widely used: Banking - loyal customer and fraud customers Medicine-Disease (patient’s medical records) Stock Market- S tock behavior, loss , Profit E-commerce- Similar customer , segmentation

16. Data Science – Modelling Method Support Vector Example : Male , Female distribution Hair Length ( c ms ) 60 40 20 0/ 140 150 160 170 180 190 200 Height ( cms )

17. Data Science – Modelling Method Support Vector Example : Male , Female distribution Hair Length ( c ms ) 60 40 20 0/ 140 150 160 170 180 190 200 Height ( cms )

18. Data Science –Model Evaluation Process Training , Test and Validation DATA Test/ Train Split Training DATA Test DATA Training Process Model Predictions

19. Data Science Demo Example

20. Demo Explanation Data 3 Species

21. Demo Explanation Load the package Caret and load the data Split the data into 2 parts -80 % would be kept in dataset and 20 % into validation Feed the dataset to 4 algorithms(CART,KNN,SV,RF) Select the best algorithm Feed the validation to best algorithm Check the output

22. Data Science Demo Installing the R platform. Loading the dataset. Summarizing the dataset. Visualizing the dataset. Evaluating some algorithms. Making some predictions

23. Other Practical's of Data Science https :// towardsdatascience.com/examples-of-data-science-with-r-789c6996435 Customer analysis and predictive analysis Association rules –( medical diagnosis, bio-medical, census data, fraud detection, CRM ) Hr Analytics - Finding valuable employees and retaining it

24. Data Science Resources Practical Data Science with R Demo commands R and R Studio installation files Resources kept at below locatio n \\gb-pb-dbm-v01\Data_Science_Resources

25. Questions and Feedbacks ?