Participated in a study measuring the Transparency and Efficiency of NCI-funded Clinical TrialsUsed RStudio to create an online dashboard detailing data sharing from clinical trials. Incorporated dynamic visualization features in the dashboard, enabling researchers and policymakers to interactively explore and understand how well the data was shared and used.Drafted a protocol for collecting publications reporting primary outcomes of clinical trials, collected these publications, and built a database to store the results.

Conducted data analysis using single-cell RNA-seq and bulk RNA-seq data across four projects focused on pulmonary fibrosis in humans and mice.Employed Fastqc and Cutadapt for data preprocessing and cleaning. Generated summary data and graphs to identify high-quality cells from samples. Utilized Seurat and Edge R for data normalization, sample integration, and downstream analyses including clustering, identification of differentially expressed genes, and pathway enrichment analysis.Created presentations to communicate findings and contributed to two publications in progress.Participated in developing a method to predict cell-cell communications using a Graph Attention Network model. Conducted three journal club to review previous methodologies, searched for applicable data, and proposed methods for testing the performance of the method.

Led a SEER-Medicare study to investigate the influence of mammography screening history on the stage at breast cancer diagnosis and mortality.Drafted a statistical analysis plan for the study; Employed multivariate logistic regression and Cox proportional hazard models in a cohort with more than 13,000 subjects and over 60 variables. Designed subgroup analyses by age and surgery status to assess result sensitivity and used statistical methods to adjust for lead time and length bias.Conducted literature reviews, drafted the study manuscript, and published it in [Journal Name]. Presented a poster of the study at the ASCO Quality Care Symposium.Participated in a study identifying factors influencing the use of broad genomic profiling using Blue Cross Blue Shield Insurance data. Conducted literature reviews, actively contributed to the team discussions on the analysis, and drafted the manuscript as the second author.

Collected pollen forecast data from commercial websites using web scraping packages in Python. Automatized the execution of the program using connectors and S3 Bucket in Amazon Web Service.Performed the data analysis using R-studio. Used Data visualization methods like residual plots, bar graphs, and scatter plots to compare the commercial pollen forecasts with empirical data.Drafted reports on the difference between forecast results and discuss factors to include to improve prediction accuracy.Presented a poster of the study to 150 participants at the Annual Health Data Science symposium at Harvard.

Used Stata and SQLite to carry out data curation for over 100,000 electronic health records, including fixing the incongruent naming of cancer stages and the diagnosis with that in ICD-O-3 and generating summary statistics.Conducted data analysis using machine learning models including random forest and logistics regression to predict cancer metastasis. Achieved an accuracy of 0.800 and an AUROC of 0.804.Performed literature review and formal analysis to summarize previous studies on using CircRNAs as biomarkers.