Published: 05 February 2018
Summary
Organizational and process pitfalls in data science and machine-learning projects have the potential to derail success and destroy organizationwide confidence in these techniques. Data and analytics leaders should use this checklist to avoid six critical project execution pitfalls.
Included in Full Research
- Prioritize Data Preparation by Allocating Sufficient Time, Staff and Funding, and Automating Whenever Possible
- Pitfall 1. Devoting Insufficient Time and Resources to Data Preparation
- Maintain a Multidisciplinary Data Science Team Equipped With the Most Compatible, Efficient Tools to Deliver Consumable Results
- Pitfall 2. Mismanaging the Data Science Team
- Pitfall 3. Employing the Wrong Tools and Building Everything From Scratch
- Pitfall 4. Failing to Interpret and Leverage Model Outputs
- Operationalize the End-to-End Analytics Process With Strong Collaboration With IT
- Pitfall 5. Failing to Operationalize Models
- Develop Ongoing Governance for Model Maintenance and Data Monitoring to Ensure That Your DSML and AI Models Produce the Correct Output Over Time
- Pitfall 6. Disregarding Model Maintenance and Data/Model Monitoring