Portfolio Project: Optimizing marketing spend using Market Mix Modeling
Go to main | Course Page
Section 1: Introduction
- Problem Statement
- Download Resources
Section 2: Understanding Data
- Packages used
- Load Data
- Data Cleaning Part - 1
- Data Cleaning Part - 2
Section 3: Feature Engineering
- Feature Engineering Part 1 - List price, payday, etc
- Outlier Treatment
- Data Preprocessing - Week finder
- Adstock variables
- EMA and SMA variables
- NPS Score
- Sales Calendar
- Climate Data
- Outlier Treatment
- Payday week and Holiday week
Section 4: Exploratory Data Analysis (EDA)
- EDA Insights Part 1
- EDA Insights Part 2
- EDA Insights Part 3
Section 5: Model Building
1 Approach 1 - Building Additive models
2. Stepwise model selection logic Theory
3. Approach 2 - Customized stepwise selection - Code Demo
4. Multicollinearity problems
5. Model building for Home Audio and Camera Accessory
6. Approach 3 - Building Multiplicative models
Section 6: Distributive Lags Models
- Outlier Treatment
- Approach 4 - Koyck Model of Distributed Lags
- Determining the net effect of Influencers - Elasticity
- Koyck Model part 2
- Approach 5 - Distributive Lag Model(Additive)
- Approach 6 - Distributive Lag Model(Multiplicative)
Section 7: Non-Linear Modeling with GAM
- Introduction to GAM modelling - Theory
- Build GAM Models Part 1 - Code Demo
- Build GAM Models Part 2 - Code Demo
Section 8: Budget Optimization Strategies
- Strategy 1 - Unconstrained Budget optimization one lever at a time (Demo)
- Strategy 2 - Budget optimization with multiple levers(Demo)
- Strategy 3 - Optimize multiple channels in presence of total budget
constraint