Maria Bulycheva

The unique challenges of data needs for research and modeling include:

• Getting sufficient coverage for the cohorts of interest
• Ensuring sufficient historical data to account for different environments, economic cycles, and other factors
• Ensuring ways to add new data sources in search of better predictability
• Having an ability to combine data from different sources (e.g., after a company merger)
• Having an ability to test predictions, correct, and improve the model capability to implement the model in production consistently with the development

Data deficiencies can have a profound effect on the quality of research and modeling forecasts, affecting the quality of the business decisions intended to be supported by the product.

This session provides examples of the challenges often faced by data and modeling teams, and how effective collaboration helps to overcome them and improve the quality of research and modeling products for the business.