Jiannan Wang (Associate Professor, SFU): DataPrep: Accelerate Data Preparation for AI скачать в хорошем качестве

Jiannan Wang (Associate Professor, SFU): DataPrep: Accelerate Data Preparation for AI

Meetup.com: https://www.meetup.com/NLP-Zurich/ Linkedin:   / nlp-zurich   Twitter:   / nlp_zurich   Summary: Data scientists have been complaining about data preparation (data collection -- data understanding -- data cleaning -- data enrichment -- data integration -- feature engineering) for many years. Although some efforts have been devoted to solving this problem, a recent survey released by Anaconda in 2020 shows that it is still the case that "Data preparation and cleansing takes valuable time away from real data science work and has a negative impact on overall job satisfaction." Most recently, Andrew Ng urged the AI community to shift from Model-Centric toward Data-Centric AI development. In this talk, I will explain what makes data preparation hard to solve, and present DataPrep (http://dataprep.ai), a fast and easy-to-use python library to address these challenges. DataPrep aims to become the "scikit-learn" for data preparation. The DataPrep library currently contains three components: a data connector component to simplify and accelerate data collection, an exploratory data analysis (EDA) component to enable fast data understanding, and a data cleaning component to clean and standardize data. I will describe their novel design and demonstrate how they can significantly save data scientists' time. In the end, I will talk about future project directions. About the Speaker: