21 April, 2017

Baidu data scientist shares how to transform data into business insights

Dr. Haishan Wu, Senior Data Scientist from Baidu Research, visited HKUST Big Data Institute on Friday, 21 April, and gave a talk on “Spatial-Temporal Brain: the spatial-temporal big data mining platform of Baidu”, sharing with audience his insights on how data could be translated to business-crucial information for the firm.

Each day Baidu handles more than 20 billion location requests from 600 million mobile users. In order to transform such large-scale data into business insights, they have developed a spatial-temporal data mining platform in Baidu called STEB (Spatial-Temporal Brain). During his talk, Dr. Wu first showed how STEB is used in user profile modeling, location based advertising, cross-device tracking, user privacy quantification and credit scoring system. STEB has also been applied in Smarter City applications, and he demonstrated two cases: chain store location selection system and human crowd forecasting system for public safety. Most recently, STEB underpins their data-driven economic measurement and investment decision system MobiMetrics. Dr. Wu also showed how MobiMetrics demystifies China’s economy and shed light on exploring with the attending audience some key questions many have in mind, e.g. where are the ghost cities in China exactly located, can we measure the unemployment rate of China, whether China is shifting its economy from investment-led to consumption driven, how does STEB accurately forecasts the Apple’s revenue in Greater China for hedge fund investors, etc., triggering heated discussions.

Dr. Haishan Wu is a senior data scientist in Big Data Lab of Baidu Research. He got his PhD from computer science department of Fudan University in China in 2011. He then joined IBM Research focusing on business data mining and analytics. Since 2012, he worked in Princeton University as a postdoc researcher. He joined Baidu in 2014 and leads a spatial-temporal data mining group. His research has been widely reported by MIT Technology Review, New Scientist, Communication of ACM, The Economist, Wall Street Journal, Bloomberg, BusinessWeek, Forbes, CNN Money, Washington Post, NPR and so on.