Data Science- An Exploratory Career

This deck was used for a guest lecture organized by the GirlScript Foundation for a session on "Data Science and It's Scope"

Transcript

2. Intro Data Science Lifecycle Why DS Role of DS Applications Learning Portfolio
3. Intro Data Science Lifecycle Why DS Role of DS Applications Learning Portfolio
4. Intro Data Science Lifecycle Why DS Role of DS Applications Learning Portfolio
5. Intro Data Science Lifecycle Why DS Role of DS Applications Learning Portfolio
6. Intro Data Science Lifecycle Why DS Role of DS Applications Learning Portfolio
7. Intro Data Science Lifecycle Why DS Role of DS Applications Learning Portfolio
8. Intro Data Science Lifecycle Why DS Role of DS Applications Learning Portfolio
9. Intro Data Science Lifecycle Why DS Role of DS Applications Learning Portfolio
10. Intro Data Science Lifecycle Why DS Role of DS Applications Learning Portfolio
11. Intro Data Science Lifecycle Why DS Role of DS Applications Learning Portfolio Benefits Low High
12. Intro Data Science Lifecycle Why DS Role of DS Applications Learning Portfolio
13. Intro Data Science Lifecycle Why DS Role of DS Applications Learning Portfolio
14. Intro Data Science Lifecycle Why DS Role of DS Applications Learning Portfolio
15. Intro Data Science Lifecycle Why DS Role of DS Applications Learning Portfolio Data Science: What’s Happening Now!
16. Intro Data Science Lifecycle Why DS Role of DS Applications Learning Portfolio
17. Intro Data Science Lifecycle Why DS Role of DS Applications Learning Portfolio
18. Data Source 1 Data Source 2 Organized Data Decision Tree Naïve Bayes Random Forest Logistic Regression Algorithms used to predict Recommend the model with highest accuracy 2006 2007 2008 . . . 2015 2016 2017 2018 Historical Data Model Building Model Testing Prediction Existing Customers Who is likely to buy? Model Results Case Study 1 – Who is likely to Buy? Intro Data Science Lifecycle Why DS Role of DS Applications Learning Portfolio
19. Decision Tree Logistic Regression Random Forest Naïve Bayes 62.8% 37.2% 46.2% 53.8% 62.1% 37.9% 61.5% 38.5% Results Algorithms Four different algorithms for prediction – Lower accuracy in each model Expected accuracy level is around 75-80% Explore other methods to improve accuracy of predicted values ▪ Data until 2015 is used for model building and predicted for 2016 ▪ The results from model were validated using actual data of 2016 Model Building Issue of Low Prediction Accuracy Intro Data Science Lifecycle Why DS Role of DS Applications Learning Portfolio
20. Intro Data Science Lifecycle Why DS Role of DS Applications Learning Portfolio
21. Intro Data Science Lifecycle Why DS Role of DS Applications Learning Portfolio