Identifying individual usage patterns for an effective promotion - - PowerPoint PPT Presentation

Sheel Jaitley – 61410271 Sunny Sapra – 61410579 Amit Phatak – 61410158 Biswajit Pattnaik – 61410498 Sourav Paul - 61410391

Identifying individual usage patterns for an effective promotion strategy

Team

Business problem

• Business problems analyzed

– Benchmark future revenue growth – Identify key customers to develop more effective promotional campaigns

• Boundary conditions

– Business segment analyzed – Point to point travel type – Time period under analysis – July 2013 to November 2013

Forecasting problem

• Forecast usage pattern for each user using logistic

regression on a weekly basis

• Each forecast indicates the user’s likelihood of

booking a cab the next week

• Data Set used

– Number of weekly booking for each user – Latitude and longitude position – User Id

Forecasting methodology

• Data visualization
• Data preparation
• Model selection
• Developing forecasting model
• Performance evaluation

Visualizing the Data Set

Data preparation and selecting forecasting model

• Filter data with correct to and from timing
• Remove incorrect data

Data preparation

• Linear regression

Model for revenue forecast

• Logistic regression

Model for Usage forecast

Model for revenue forecast

• Predictors:

– Lag 1 to lag7 for number of daily booking

Forecasting model

• Y= β0 + β2*Wt-1 + β2*Wt-2 +β2*Wt-2 …….β2*Wt-7

Performance Evaluation

Residuals Actual vs Predicted vs Naive

Forecasting weekly usage pattern for each user – single user model

• Predictors:

– Number of booking per user in the week (W) – Lag 1 and lag 2 of weekly bookings of the particular user

Forecasting model

• Logit(week=1) = β0 + β1*Wt-1 + β2*Wt-2

Performance evaluation

Output for naïve forecast Output for logistic regression

Classification Confusion Matrix Predicted Class Actual Class 1 1 3 3 Error Report Class # Cases # Errors % Error 1 3 3 3 100 Overall 6 3 50 Classification Confusion Matrix Predicted Class Actual Class 1 1 1 2 2 1 Error Report Class # Cases # Errors % Error 1 3 2 66.67 3 2 66.67 Overall 6 4 66.67

Forecasting weekly usage pattern for each user – single model for multiple users

• Predictors:

– Number of booking per user in the week (Wn) – Lag 1 and lag 2 of weekly bookings per user

• Interaction variable

– D1, D2 … Dn for n users

Forecasting model

• Logit(week=1) = β0 + β1* D1*Wt-1 + β2* D1*Wt-2

+ β3* D2*Wt-1 …………… β2n* Dn*Wt-1

Performance evaluation

Classification Confusion Matrix Predicted Class Actual Class 1 1 11 8 8 Error Report Class # Cases # Errors % Error 1 19 8 42.11 8 0.00 Overall 27 8 29.63

Output for naïve forecast Output for logistic regression

Classification Confusion Matrix Predicted Class Actual Class 1 1 6 4 13 4 Error Report Class # Cases # Errors % Error 1 10 4 40.00 17 13 76.47 Overall 27 17 62.96

Performance evaluation