Tianyi Shi Email & Phone Number

Alvin, Texas, United States

• Train a Cascade Classifier to extract maps from day-to-day energy-related pdf files (40K+) to keep the industry information up to date and save time for stakeholders• Detect data tables with various alignments via CascadeTabNet, Tabula and Camelot from scanned PDF documents, and obtain an abstract of the table by keywords for the president to review

Champaign, Illinois, United States

• Established a predictive model based on applicant admission of the University of Illinois data (92146 rows with 256 features) using machine learning libraries, such as scikit-learn, XGBoost and statsmodels, in Python• Helped the admission office to better target applicants who are likely to accept offers and enroll in order to improve yield rates and save manual reaching-out efforts

Champaign, Illinois, United States

• Executed MySQL LAG function to compute hourly utility usage of Year 2023 for university buildings and harnessed Power BI to create time series charts for 7-day cycle trend analyses, enabling precise tracking of periodic patterns. Developed an alert system in Power BI to monitor real-time utility usage against historical averages, triggering alerts when usage exceeded 30% above average, ensuring proactive energy management• Validated data integrity of university enrollment demographic data over past 20 years in Excel by addressing null values, handling duplicate values and conducting a sense-check; Visualized the ethnicity and gender distribution using pie charts, stacked bar charts, and line charts in Tableau to assure the improvement in student diversity• Applied web scraping technique to collect IT Leadership Workshop alumni data from 2010 to 2023; Leveraged an LLM model in Python to consolidate varied unit names into a consistent format to optimize the readability for future data process, analysis and modeling

Champaign, Illinois, United States

• Designed and synchronized Instagram and Facebook campaigns, strategically timing posts for optimal engagement and maximizing reach for a 10% increase in daily active users• Integrated Google Analytics to monitor social media content performance, leading to data-driven strategies for improved engagement by 15%

Suzhou, Jiangsu, China

• Employed pandas of Python for data cleaning of raw databases containing 10K+ rows to enhance the accuracy and security measures, including removing missing values, merging multiple datasets, and standardizing the format• Utilized Vlookup in Excel to classify and report non-standardized software based on cleaned data, and worked as a cross-function team member to address the problem, achieving a 55% reduction in company’s data leakage risk • Created Power BI visualization reports to depict employees' compliance of cyber security, and implemented data-driven decisions to enhance data security level by 45% for 7000+ employees

Bachelor Of Science - Bs, Statistics

High School Diploma