Data Science Scholar – completed 3 projects with 10 DSs leveraging AI/ML in Azure cloud environment1. Data Pipeline Deployment for customer analysis with NLP model:• Conducted sentiment analysis and topic modeling analysis on over 100,000 customer feedback utilizing BERTopic, Roberta and other Python packages with 3 DSs in the team• Built a weekly data pipeline in Databricks to automate functions, including data cleaning and modeling, delivering, and displaying insights on a Power BI dashboard in Azure Explorer.• Integrated Azure ChatGPT into Power Apps to enrich the reporting dashboard interactivity and increase the efficiency of customer feedback analysis.2. Visualization trial for data-driven customer service improvement:• Collaborated with 5 DSs and developed an interactive dashboard in Power BI covering visualizations of insights fromsentiment analysis, topic modeling analysis and word association networks.• The dashboard is used to effectively assess the performance of service centres in NSW and support the decision-makingprocess on continuous service improvement.3. Predictive Model for Topic Modelling:• Collaborated with 2 DSs to develop a predictive model that enhances topic modeling performance by fine-tuning sentence transformers using the Setfit framework.• Lift the F1 score for the model from 0.58 to 0.71 and proved an efficient topic modeling method to be utilized for future text-related projects.

1. ETL:• Extracted, cleaned, and wrangled logistics route data from APIs using Python, laying the groundwork for navigation evaluation.2. Model Tuning:• Identified and rectified incorrect map and logistics route data from various relational datasets using Python and SQL• Analyzed the bias of the navigation algorithm and visualized results by bar plot and histogram to inform algorithmic improvements.3. Documentation:• Supporting the development of a new navigation algorithm by summarizing and documenting findings from research papers.