Chen Song Email & Phone Number

Los Angeles, Ca, Us

Los Angeles, California, United States

• Lead and design the end-to-end development of the PnL PowerBI dashboard that is used by the Executive Team on a daily basis, reducing the inefficient manual work and increasing the report accuracy.• Optimized SQL queries across dashboards to enhance computational efficiency and speed.• Identify data gaps in the analytics teams internal database and collaborate with engineers to ensure gaps are resolved to meet the businesses needs.• Provide ad-hoc data supports using SQL query and Excel to fulfill stakeholders blindspot and answer their key questions.

Berkeley, California, United States

• Loxz Digital Excellence in Data Science Award Oct 2021: Led and developed an end-to-end machine learning solution for the Email Campaign Optimizer for marketing email click-through rate. Fine-tuned a pre-trained BERT model for industry classification to impute missing industry labels.• Built and maintained Tableau analytical dashboards to monitor the KPI metrics and provide business insights.• Designed and created an ETL data pipeline of sales and orders with Snowflake and AWS, successfully reducing manual work by 40%.• Led survey data analysis and methodology report capturing vital business insights by analyzing and structuring 2M+ raw data, and visualizing data using Tableau. Support product launch decisions on the Loxz ML Readiness survey.Advanced tech skills & tools: AWS, S3, Snowflake, PostgreSQL, Classification, Tableau, NLP, Python

Berkeley, California, United States

• Loxz Digital Excellence in Data Science Award May 2021: Completed survey data analysis and methodology report capturing vital business insights by identifying 20+ features and leveraging text mining techniques on respondent data.• Introduced and managed a peer-review framework using Git and Colab, improving teamwork efficiency and documentation by 15%.• Presented 7+ analytical and visual reports to potential funding government organizations and successfully gained partnership with UCI.

Washington Dc

• Developed a web application for Ella using Python and Streamlit, successfully improved webpage load speed by 10% by optimizing steamlit caching process.• Enhanced Snorkel model performance by 20% by adding 3 customized text features to the Snorkel classification model.• Completed NLP review summarization on 50K+ customer reviews of (Dubai) Zomato App and generated interactive dashboards in Tableau, aiming to provide insights of customer behavior changes and vital reviews to stakeholders.

Greater Seattle Area

• Created an end-to-end Machine Learning solution analyzing people’s reactions to popular events using NLP techniques.• Proposed 4 definitions of the absolute volume of google trends and successfully improved the model accuracy by 27%.• Merged Emoji analysis to text analysis which improved model performance by 5%.• Developed Naïve Bayes, XGBoost, RandomForest, LassoLars and SVR models, tuned the hyper-parameter and successfully beat the baseline model.

Seattle, Washington, United States

• Successfully improved waste collection efficiency by 40% by implementing ML models and generating trash can buffers using GIS data.• Solved “lack of ground truth” data issue by proposing a cost function of collecting trash with different trash levels.

Master'S Degree, Information Management, 3.8 Out Of 4.0

Coursework: Natural Language Processing, Data Science, Business Intelligence Systems, SQL for Data Science, Advanced Database, Computer.

Bachelor Of Management In Public Utility Management, Public Administration