Manqiong Chen

• Assisted approximately 7-8 students for statistics based university programs, reviewing course material and ensuring knowledge was properly acquired through practice and exam preparation.
• Reviewed previous assignments and exams to find areas of improvement for students.

• Assisted professor within a university level statistics lab, helping students acquire proficiency in a variety of programs including JMP, R and both modular and statistics programs. Assisted students by responding to any additional questions they might have outside of classroom scope through weekly office hours.
• Conducted structured review sessions for students, summarizing classroom knowledge and principles. Operated as a point of contact for students, assisting with solving any statistical and software problems.
• Aided professors through grading and evaluation of student assignments and examinations.

• Design and develop state-of-the-art machine learning models and algorithms to detect and prevent fraudulent activities, such as account takeover, card takeover, and transactional fraud.
• Collaborate with cross-functional teams, including data scientists, software engineers, and product managers, to define the requirements and scope of fraud detection projects and develop solutions that align with business goals.
• Conduct extensive data analysis, feature engineering, and data preprocessing to extract meaningful insights from large-scale, complex datasets to improve fraud detection accuracy and efficiency.
• Evaluate, select, and implement appropriate machine learning techniques, algorithms, and frameworks, such as anomaly detection, and ensemble methods, to enhance fraud detection capabilities.

• Building and testing up-sell and cross-sell recommendation models, such as ALS Collaborative Filtering Model, CLTV model by using pyspark and scala spark.
• Maintenance of and optimizing the current existing trend filtering model.
• Building streamlit dashboard to handle model deployments and experimentation.

• Develop the data product by using pyspark and python.
• Work closely with MLE for machine learning model development (that includes interpreting machine learning model predictions, model validations).
• Collaborate with business users to develop new data products.
• Prepare and automate dashboard from statistical analysis for review include graphs, tables, and interpretations.

• Access, analyze, and interpret customer transactional data on behalf of Precima clients to drive insights using statistical techniques in support of business objectives such as category management, marketing strategy development and execution, and ad hoc strategic and tactical client issues.
• By using statistical and machine learning methods to build the B2B Segmentation Model, then test, validate, and automate model.
• Responsible for Analytics production support and deployment, which includes executing, developing, deploying, and measuring Precima's analytical products that are configured for clients.

• Performed A/B Testing to find differences between customer promotions. Where there was no difference, we did investigate the flaws in the experimental design.
• Qualified the value of time for each customer. This would be inputted into a price elasticity model to discover the upper bound threshold for increasing toll rates before losing revenue.
• Collected and aggregated data from internal and third-party sources to build profiles of our customers so we could predict and influence future behaviors and outcomes.
• Automatically collected collision/construction data on major and arterial roads. Data used as input to our price elasticity models and to provide customers real-time updates of road congestion (NLP).
• Predicted and understood the key drivers that impact the transponders’ lifetime duration (Survival Model).

Digital Marketing Project

https://manqiong.wordpress.com/2017/10/23/google-analytics-reporting- project/

• Goal of project was to find out how likely a user would access WeCloudData Web to help the company to raise its prestige.
• Performed various Chi-Squared Goodness-Of-Fit (Scipy) tests on data collected from Google Analytics Reporting API.

Data Science House Price Prediction Project

https://manqiong.wordpress.com/2017/10/22/data-science-house-price- prediction-project-2/

• Developed a model to describe and predict house prices based on data collected from the web. Features included house size, neighborhood safety, number of rooms, traffic, etc.
• Scraped web data from Kijiji using NLP toolkit (NLTK).
• Constructed various regression models (sklearn): linear regression, lasso regression, SVM, and decision tree.
• Used cross-validation for model selection.

• Constructed a new method for minimizing the variance of the least-square estimates of the parameters in a linear model. Discussed two types of directional derivative and some links between them.
• Addressed different optimal criterion such as D-optimality, A-optimality, etc. by deriving derivative criterion and solving the equations.
• Constructed design using the class of algorithm and improving the convergence of the algorithm by the property of the directional derivative.
• Give presentation to all professors from different departments.