1. Nudge
1.1. Nudge model
1.1.1. Parameters
1.1.1.1. Freq
1.1.1.2. Time
1.1.1.3. Type
1.1.1.4. Location
1.1.1.5. Day of week
1.1.1.6. Weather
1.1.2. Evaluation
1.1.2.1. App activity (feedback loop)
2. Library Optimizer
2.1. Most engaging personalized parameters
3. Platform Requirements
3.1. Early phase
3.1.1. Heuristic Approaches
3.1.1.1. Data collection
3.1.1.1.1. Insight
3.2. Mature phase
3.2.1. Sufficient data collected
3.2.1.1. Learning-based approaches
3.3. A/B Testing
3.3.1. Canary Release
4. Feedback loop
4.1. Implicit
4.1.1. Activity / achievement
4.2. Explicit
4.2.1. Thumbs up / like
5. Questions
5.1. Other channel to reach the users?
5.1.1. Feedback loop for different channel response?
5.2. Existing data for "engagement" effectiveness?
5.3. How many users in total?
5.4. How many "Daily Active Users?"
5.5. Current user profile, bias in distribution?
6. engagement
6.1. incentive
6.1.1. user targeting
6.1.2. actions
6.1.3. result
6.1.4. Example
6.1.4.1. Drink more water
6.1.4.1.1. Encourage drinking more water
6.2. Social
6.2.1. user goruping
6.2.1.1. user similarity
7. User targeting
7.1. Begin from heuristic
7.2. Automatic targeting
7.3. Recommender
7.3.1. Challenge embedding
7.3.2. User embedding
8. Data Types
8.1. User Info
8.1.1. age
8.1.2. gender
8.1.3. height/weight/bmi
8.2. Sensor Data
8.2.1. heart rate
8.2.2. Blood pressure
8.2.3. sleep
8.2.4. Active hours
8.2.5. Lifestyle (walking, sitting, driving, cycling)
8.3. Medical/Clinical data
9. DS model
9.1. Nudge model
9.1.1. Parameters
9.1.1.1. timing
9.1.1.2. location
9.1.1.3. channel
9.1.1.4. environment
9.1.1.4.1. indoor
9.1.1.4.2. weather
9.2. Recomnender
9.2.1. Early: heuristic (already there)
9.2.1.1. Efficient for "cold-start" situation
9.2.2. Later: learning-based model
9.2.2.1. targeting
9.2.2.2. engagement prediction
9.3. Challenge library
9.3.1. challenge embedding
9.3.2. challenge optimizer
9.3.2.1. search for optimal parameters of a challenge for an individual
9.4. User embedding model
9.4.1. user similarity / grouping
9.4.2. Life style model
9.5. Incentive: Social Engaging model
9.5.1. user grouping
9.5.2. group challenge
9.6. architecture
9.6.1. Transfer learning
9.6.2. Multi task learning
9.6.3. embeddings
10. DS Platform
10.1. Data Platform
10.1.1. Who will be the users?
10.1.1.1. access / permission control
10.1.1.2. User account system design
10.1.2. Any existing access/permission control system?
10.1.2.1. Is it a service can be used for authentication by other system?
10.1.3. Desensitization level?
10.1.4. Latency to update libraries/models?
10.1.4.1. Real-time or batch
10.1.4.1.1. Data lake design
10.1.5. Challenge library management?
10.2. Experiment Management Platform
10.2.1. Library experiment management
10.2.1.1. version control?
10.2.2. Nudge management
10.2.3. Gamification platform
10.3. MLOps platform
10.3.1. online learning
10.3.2. Approval before launch
10.3.3. canary release?
10.3.3.1. who's the mobile app developer?
10.3.3.1.1. HTML5? Need release for new feature?
10.3.4. Do they have SRE team?
10.3.5. MLOps
10.3.6. Model release
10.3.7. Model deployment
10.4. BI / DA Platform
10.5. A/B platform
10.6. AWS / Azure
10.7. Program Library Management
10.7.1. Nudge
10.7.2. Challenge