Demographics Analysis

For gender, age group, income level, etc

Online behavior Analysis

For recency, frequency and monetary signals of audience

Interest and Life Style Analysis

Identifying audience's interests for right campaign such as Sports, Movies, Shopping

Location Analysis

Where they live/work, places visited and travel patterns and predictions

Campaign Response Analysis

Real-time and historical campaign performance analysis for segment optimization

Audience Insight Reports

Integrate all types of audience analysis into insight dashboards and reports for pre/mid/post-campaign analysis

Cross device ID Mapping

Probabilistic and Deterministic approaches for the association of device IDs, to the same person

Machine learning enabled user profiling

- Google BigQuery + Apache Spark
- Covariance Matrix
- Decision tree based models and support vector machine based models + Input Space + Feature Space

Feature Generation
Clean and encode data (raw data -> tons of features: X')

Feature Section
Select features based on the correlation between X' and Y (X'->X)

Model Training
Train models base on feature characteristics (X,Y)->{f1(x),...,fn(X)}

Model Selection
Select / ensemble models {f2(x),...,fn(x)}->f'(x)