Online Program

Project Development and Evolution at AT&T Labs: Iterative Research in Practice (309621)

Abstract: As AT&T’s research and development division, AT&T Labs has a notable history for important contributions to the theoretical and applied sciences, and scientists continue to conduct exciting research at the intersection of statistics and machine learning, optimization, and computing. Problems are often motivated by internal business units as diverse as mobility network engineering, content and advertising services, and retail store operations. In this talk, Emily will discuss the development of a recent project with AT&T’s Federal Regulatory organization to analyze emergency call data, an investigation that took her to Monash University in Melbourne, Australia for an academic research collaboration with Professor Di Cook to develop methods for interactive visualization of multivariate spatiotemporal data. The iterative nature of data science research necessitates flexibility to move through three distinct but not independent technical project phases—problem statement and question definition, approach and methodology, and computational implementation—because the limitations from one phase, often related to scalability, may inform choices made in another. Emily will present the technical contributions of her project and demonstrate its evolution as business partners and mentors shared their expertise throughout this data science process.