Leveraging Visual Analytics for Data Science

Visual analytics offers a wide range of techniques to explore large and complex datasets. In this talk, the speaker will explain how some of these techniques can be leveraged to develop superior statistical models to address common machine learning driven data science use cases such as large-scale classification - a problem commonly faced in the e-commerce (for categorizing commodities into several categories) and many other domains. The talk will guide the audience through engaging case studies to show how interactive visualizations, tweaked and twisted to suit the scenario, can lead to better understanding and modeling of the data. The talk will highlight the role of visual analytics at every stage: understanding the data at hand, extracting features, building a model, evaluating it, performing model diagnostics, and comparing multiple models trained for the same task.

Abon Chaudhuri develops machine learning driven solutions for product content enrichment and product categorization at @WalmartLabs, Silicon Valley. Earlier, he worked on analyzing and visualizing nano-scale imaging data at Intel Corporation. He has experience of visualizing scientific data produced by supercomputers at Argonne and Oak Ridge National Laboratory. Abon received his PhD in Computer Science and Engineering from The Ohio State University in 2013. His doctoral research focused on feature-driven summarization of big data to enable visualization and query-driven analytics. Abon has authored several research publications that earned him three best paper/poster awards and a visualization award at top-tier IEEE conferences.