Zijie(Scott) Liu

Experience

• 

  The Wright Star LLC

  Oct 2019 - Jan 2020

  Data Analyst

  • Partnered with a mentor to analyze P/E Ratio Model, utilized MS Excel to classify scale score of ratio value, determined value of ratio varies by industries, defined range of values group by each industry• Built classification model using Python, input daily stocks file and appended to analyzable sheet by sectors, calculated outliers boundary, number, percentage in all observations, described quartile values of different ratios• Optimized classification model and fixed errors in daily analysis, reduced 80% of time and nearly 20% of deviation in prediction process, increased around 15% on average of customers’ final potential payback on stocks investment Show less

• 

  Cloud9 Advisory, Inc

  Oct 2021 - Dec 2021

  Data Analyst

  • Preprocessed over 11 million records and 54 variables of Paycheck Protection Program (PPP) Data and assisted inmaintaining the ETL pipeline in Python, successfully classified the data by business type and industry under anautomated R solution to save an average of 20 labor hours weekly and improved data collection quality• Aggregated the processed PPP datasets, delivered the EDA package to automatically generate useful insights fromraw records, and visualized forgiveness amount data by locations and dates in dashboards using Tableau to keepmonitoring the Second Draw PPP Loan data, increased the exploring speed of the whole dataset by 50%• Built multiple regression models to predict the forgiveness results, optimized the model accuracy by using MachineLearning techniques (Logistic Regression, Decision Tree, Linear Regression, Cross-validation), and presented the overall recommendations & forgiveness status naturally to stakeholders, tripled the efficiency of Agile methodology Show less

• 

  KAYMILE TRADING INC

  May 2022 - May 2023

  Data Analyst

  • Gathered sales, production, and inventory data, aggregated the data in Python to ensure data collection consistency• Created dashboard in Power BI, transformed the data using Power Query Editor for creating monthly report• Used historical data to develop Demand Forecasting Models (Time-series & Regression Analysis, RecommendationEngine, TF-IDF, NLTK) for various products, predicted future sales trends, reduced inventory waste by 7%, improved overall accuracy of inventory & import demand by 18%, promoted data-driven decision-making across the team Show less

• 

  Joblogic-X Corporation

  May 2022 - Feb 2023

  Data Scientist

  • Preprocessed millions of records of bank loan data, accomplished database normalization (Designed database schema, Extracted and Transformed data, Testing and Verification migration), delivered EDA packages• Implemented the Cross-validation on model selection to predict Credit Risk in both supervised and unsupervised learning, optimized the AUROC score of the model classifier, and improved an average of 12% model performance• Built Meetfresh Recommendation Engine based on indicators, deployed Content & User based Analysis by using NLP techniques (Text Preprocessing, TF-IDF, NLTK), provided user feedback analysis to improve business performance Show less

• 

  GTSP Group

  Apr 2023 - now

  Data Scientist

  • Managed and optimized a 620GB PostgreSQL database for InstaGrow, designed complex SQL queries for data analysis, and collaborated in the DevOps lifecycle. Integrated data pipelines using Informatica, and achieved an average 15% increase in user engagement (daily new users, follow back rate, session duration)• Processed EVE Online database by using Sqlite3, crawled the data by accessing the EVE market XML API endpoint, created an automated market transaction monitoring tool in Python, saved an average of 2 labor hours daily• Converted the raw data of corn production and precipitation from USDA, aggregated the data to determine the US top producer, correlated variable relationships, and built the statistical model (Linear Regression, SARIMA, SVM) in Python to predict the future agriculture production Show less