Presented By O'Reilly and Cloudera
Make Data Work
22–23 May 2017: Training
23–25 May 2017: Tutorials & Conference
London, UK

Mister P: Imputing granularity from your data

Rumman Chowdhury (Accenture)
14:0514:45 Thursday, 25 May 2017
Level: Intermediate
Average rating: ***..
(3.00, 1 rating)

Who is this presentation for?

  • Data scientists

Prerequisite knowledge

  • Familiarity with basic statistics and survey analysis

What you'll learn

  • Gain an overview of multilevel regression and poststratification (MRP)
  • Learn how to derive more granular data from general data


Multilevel regression and poststratification (MRP) is a method of estimating granular results from higher-level analyses. While it is generally used to estimate survey responses at a more granular level (e.g., by academics in political science to estimate state-level opinion from national polls), MRP has clear applications in industry-level data science. Rumman Chowdhury reviews the methodology behind MRP and provides a hands-on programming tutorial.

Topics include:

  • A review of the statistical concepts behind MRP and how it works
  • A comparison of MRP to other similar methods
  • Benchmark data on performance
  • An example of MRP in practice (imputing county-level public opinion about government using national-level polls)
  • Translating MRP to industry
Photo of Rumman Chowdhury

Rumman Chowdhury


Rumman Chowdhury is a senior manager and AI lead at Accenture, where she works on cutting-edge applications of artificial intelligence and leads the company’s responsible and ethical AI initiatives. She also serves on the board of directors for three AI startups. Rumman’s passion lies at the intersection of artificial intelligence and humanity. She comes to data science from a quantitative social science background. She has been interviewed by Software Engineering Daily, the PHDivas podcast, German Public Television, and fashion line MM LaFleur. In 2017, she gave talks at the Global Artificial Intelligence Conference, IIA Symposium, ODSC Masterclass, and the Digital Humanities and Digital Journalism conference, among others. Rumman holds two undergraduate degrees from MIT and a master’s degree in quantitative methods of the social sciences from Columbia University. She is near completion of her PhD from the University of California, San Diego.