数据科学家:岗位职责、能力要求与人才培养

Typical Responsibilities, Key Qualifications and Higher Education for Data Scientist

数据科学家具有区别于其他专业人才的能力要求和岗位职责,培养数据科学家是数据科学与大数据技术专业的主要使命。本文以Indeed、LinkedIn和百度百聘为数据来源,广泛搜集中国、美国、英国、德国、加拿大、日本、澳大利亚和韩国八个国家五种语言的数据科学家招聘公告,挑选出206则具有代表性的招聘公告,对其中的任职资格要求和岗位职责描述进行聚类分析和观点挖掘,提炼出数据科学家的能力要求和岗位职责。经调查发现,数据科学家的主要岗位职责有:以数据为中心提出解决方案,从海量数据中洞察有价值的信息,面向具体业务的算法模型研发,假设检验与试验设计,数据治理与数据质量控制,数据产品的研发及基于数据的传统产品的创新,数据全流程的参与以及跨部门和跨领域合作等;数据故事化、因果分析、实时流式处理、部署/生产模型等新兴业务需求将成为未来数据科学家岗位职责的新增长点;数据科学家与数据科学相关的主要能力要求包括:SQL编程、Python/R/SAS、Hadoop MapReduce/HBase/Hive、Spark/Storm、基于Tableau等的可视化分析、ETL处理、数据仓库/数据湖/BI技术、统计学与机器学习(含深度学习)、自然语言处理及文本分析和机器视觉等。图4。参考文献21。

A survey for collecting data scientist job announcements from Indeed, LinkedIn and Baidu Baipin is conducted, and 206 typical cases are selected for the study, which involves 5 languages and 8 countries, including China, the United States, the United Kingdom, Germany, Canada, Japan, Australia, and South Korea. Then, the key qualifications as well as typical responsibilities of data scientists are described via utilizing cluster analysis and opinion mining to provide a basis for the training of data scientists, especially the construction of data science and big data technology major. The qualifications for data scientists can be divided into two categories: data science-specific qualifications and general purpose-oriented ones. Data science-specific qualifications include SQL programming, Python/R/SAS, and Hadoop MapReduce/HBase/Hive, Spark/Storm, Visual Analysis with Tableau, ETL, Data Warehouse/Data Lake/BI, Statistics, Machine learning(including deep learning), Natural Language Processing, Text Analysis, and Computer Vision. General purpose-oriented qualifications mainly involve the candidate’s readiness of communication and cooperation, problem-solving, 3 C characteristic of data scientists, independent learning, attention to detail, stress management, and leadership skills. The main responsibilities of data scientists include designing data-centric solutions, finding valuable insights from massive data, developing algorithms/models for specific businesses, hypothesis testing and experimental design, data governance and data quality control, R&D of data products, the innovation of traditional data-based products, as well as participation in the whole data process, cross-department/domain cooperation. Besides, personal charisma, experiences of participating in big data competitions and open source communities, the quality of full-stack data scientists, mathematics and programming capabilities, user-centered design methods, and humanistic issues including big data privacy protection, have an important influence on the core competencies of data scientists. At the same time, emerging business requirements such as data storytelling, causality analysis, real-time flow processing, and deployment/production model, will become novel topics of emerging qualifications of data scientists in the future. The main implications of this study for the data science and big data technology major in China are to focus on the curriculum of data science itself, to introduce Industry-University-Research cooperation, to promote the theoretical research of data science, to develop several core courses, to leverage the capstone role of data product development in data science, to help students develop their self-learning skills, and help students master the basic knowledge and skill of data scientists. Data scientists’ key qualifications and typical responsibilities are unique, and to cultivate data scientists is the main mission of data science and big data technology major. 4 figs.21 refs.