• Metrics Dependency Graph: Used Scope to get outlook search data, then used CausalDiscoveryToolbox in Python to extract causality between different metrics, then used Neo4j to construct an interactive Metric Dependency Graph which allows users to filter, search, find root cause and export. • Metric Design: Investigated calculation logic of metrics and created a rule-based metric, which was adopted.

• Project Foundation: Got questionnaire and customer information data from a telecommunication company. Used these data to get user segmentation for further personalized marketing and found important features.• Project Workflow: Used Max-Diff to get utility score and consensus clustering to get user clusters via R. Used Over-Indexing analysis to define each cluster. Used tree models to get important features via Python; Iterated among feature combinations, encoding methods and algorithms;… Show more • Project Foundation: Got questionnaire and customer information data from a telecommunication company. Used these data to get user segmentation for further personalized marketing and found important features.• Project Workflow: Used Max-Diff to get utility score and consensus clustering to get user clusters via R. Used Over-Indexing analysis to define each cluster. Used tree models to get important features via Python; Iterated among feature combinations, encoding methods and algorithms; Got robust results via Chain classification. Show less

• Basic work: Used SQL to get data and output daily/weekly/monthly reports, monitored key indexes and analyze abnormal cases• Data board construction: Constructed 5 data boards of detail page and message center of Tencent Video. Created data sources and finished data paths• Experiments: Conducted 3 experiments such as the effect of using fake teasers of TV series and output reports

• Project description: Predicted the demand of Qingju Bicycle, found abnormal factors, and quantified the impact to improve the matching efficiency of supply and demand. Formed a new product design• Team work: Was responsible for data preprocessing, feature construction, data visualization, panel data regression, ML prediction

• Data support: Provided corresponding data using Hive to make a better decision; Output weekly reports• Data analytics: Analyzed the review of top active families in February and other 4 topics and output reports; Built author evaluation index systems to assist business decision-making • Data monitoring: Participated in constructing an index system of PK function to monitor performances• Research Projects: Explored the effect of user follow behavior on authors, and output the report

• Project management: Coordinated with product managers, R&D team, and financial department; Participated in the whole process of “Talent Writing” project and output daily and weekly reports • Meetings: Organized 4 kinds of meetings such as project approval meetings, weekly meetings, and output meeting minutes• Work optimization: Wrote project management specification manuals and project manager operation manuals; Initiated and maintained different projects on the PMP system

• System DevOps: Completed daily fund clearing system running batches, Baidu Baiying Fund sales platform fund product maintenance and synchronization docking and other basic work, clearly grasped the operation process of fund company and fund agency platform• Optimization: The member of the fund clearing system optimization team, proposed opinions such as setting default days• Analysis and summary: Provided detailed operation procedures and problem solutions for running batches, wrote… Show more • System DevOps: Completed daily fund clearing system running batches, Baidu Baiying Fund sales platform fund product maintenance and synchronization docking and other basic work, clearly grasped the operation process of fund company and fund agency platform• Optimization: The member of the fund clearing system optimization team, proposed opinions such as setting default days• Analysis and summary: Provided detailed operation procedures and problem solutions for running batches, wrote operation manuals and special situation handling manuals which were adopted Show less

• Product design: The member of the doctor-side medical records collection paid function team, participated in the design of user-side medical record framework and the details of background• Improve user satisfaction: Selected four competing products from the perspective of interface experience, completed competitive analysis report, proposed opinions such as enriching search options• Data analysis: Used Python to analyze physician data, visualized the regional distribution, evaluation,… Show more • Product design: The member of the doctor-side medical records collection paid function team, participated in the design of user-side medical record framework and the details of background• Improve user satisfaction: Selected four competing products from the perspective of interface experience, completed competitive analysis report, proposed opinions such as enriching search options• Data analysis: Used Python to analyze physician data, visualized the regional distribution, evaluation, popularity of physicians, and updated homepage ads monthly; Created user questionnaire and published it on the platform, which had thousands of user feedback Show less