Successfully reported this slideshow.
We use your LinkedIn profile and activity data to personalize ads and to show you more relevant ads. You can change your ad preferences anytime.

Demo: Using Azure Machine Learning to Predict Churn

47 views

Published on

Learn how you can leverage the elastic, on-demand processing power of Microsoft Azure to create faster, more applicable analytics. Data Scientist and Author, Ahmed Sherif, will demonstrate key analytic use cases that can be spun up quickly with minimal effort and maximum return on investment.

Published in: Technology
  • Be the first to comment

  • Be the first to like this

Demo: Using Azure Machine Learning to Predict Churn

  1. 1. Azure ML Churn Demo From Scratch November 2nd, 2017 Twitter: @theAhmedSherif
  2. 2. Review our Customer Case Demo #1: Importing Data Demo #2: Identifying Missing Values Demo #3: Building Logistic Regression Model Demo #4: Comparing 2 Scored Models side by side Demo #5: Notebooks and Web Services
  3. 3. Review Our Customer Case
  4. 4. Keep Calm and Minimize Churn A large multimedia conglomerate is concerned customers are leaving for more online media content Cord Cutting!!!! They need to understand who is leaving and what factors are indicators for churn You work for this Company! You need to help them figure out what kind of customers are leaving and for what reason
  5. 5. Dataset 7,000 records The following columns customerID Online Backup gender Device Protection SeniorCitizen Tech Support Partner Streaming TV Dependents Streaming Movies tenure Contract PhoneService Paperless Billing MultipleLines Payment Method Internet Service Monthly Charges Online Security Total Charges
  6. 6. Modeling Strategy The Target Column is Churn All other variables are Predictors The output is Binary: Yes or No for Churn Ideal Candidate for binary classification model Logistic Regression Decision Forest
  7. 7. Demo #1 Importing Data and Exploring Fields in Azure ML
  8. 8. Demo #2 Identifying Missing Values and Imputing Results
  9. 9. Data Cleansing 80% of the time a Data Scientist is spent cleaning dirty data 20% of the time a Data Scientist is complaining about cleaning dirty data
  10. 10. Demo #3 Building Logistic Regression Model, Scoring, and Evaluating
  11. 11. Demo #4 Comparing Logistic Regression and Decision Forest Models
  12. 12. Demo #5 Notebooks and Web Services
  13. 13. What about Bob?
  14. 14. Questions?
  15. 15. THANK YOU! What questions do you have?

×