Wednesday, 11 March, 2020 01:45 PM|Wednesday, 11 March, 2020 02:30 PM
Drive Data Scientist Productivity With Data Wrangling Techniques
Sumit Agarwal, Sr Director Analyst, Gartner
Anywhere between 40% - 80% of a data scientist's time is spent in data analysis, and finding the right data alignment with the machine learning algorithm. The current ad hoc and iterative nature creates uncertainty in the project implementation timelines, as well as impact the model output quality and fairness in decision making.
This session provides various techniques such as using specialized tools, crowd-sourcing, simulations and synthetic data creation, and use of generative adversarial networks (GANs) to substantially improve the model build process using wide variety of data such as transactional, text, images, videos, speech etc.