• Introduced BeautifulSoup for web scraping news headlines from Yahoo Finance over the past five years, social media tweets related to Tesla from X, and Tesla earnings calls as data sources.• Cleaned all data by removing stop words and irrelevant terms.• Integrated an LLM model pipeline for pre-training sentiment analysis on the dataset.• Fine-tuned the BERT model to create our sentiment score dataset.• Utilized Pandas and Numpy for data cleaning and preprocessing on the… Show more • Introduced BeautifulSoup for web scraping news headlines from Yahoo Finance over the past five years, social media tweets related to Tesla from X, and Tesla earnings calls as data sources.• Cleaned all data by removing stop words and irrelevant terms.• Integrated an LLM model pipeline for pre-training sentiment analysis on the dataset.• Fine-tuned the BERT model to create our sentiment score dataset.• Utilized Pandas and Numpy for data cleaning and preprocessing on the sentiment score dataset, including splitting the data into training and validation sets.• Connected to Gemini's API for data labeling and clustering, categorizing Tesla into different industry sectors.• Incorporated financial report data and fed all parameters into an LSTM model for stock price prediction.• Achieved a Loss (MSE) of 9.73, IRR of 12.83%, and Winning% of 62.38% with our stock price prediction model. Show less

• Responsible for analyzing the past 10 years of project data for the Development Foundation, aiming to create a comprehensive database and manage it efficiently.• Cleaned project names data and used CKIP Tagger's LLM for word segmentation to form a vocabulary dataset.• Utilized Pandas to cluster the dataset based on similarity and labeled the clusters.• Successfully consolidated over 80% of the database project names, established a complete SOP, and aimed to predict future project… Show more • Responsible for analyzing the past 10 years of project data for the Development Foundation, aiming to create a comprehensive database and manage it efficiently.• Cleaned project names data and used CKIP Tagger's LLM for word segmentation to form a vocabulary dataset.• Utilized Pandas to cluster the dataset based on similarity and labeled the clusters.• Successfully consolidated over 80% of the database project names, established a complete SOP, and aimed to predict future project development trends. Show less

• Collaborated with 50 team members to coordinate various aspects of the event and prepared for a camp with over 1000 participants while maintained partnerships with multiple companies, enhancing event sponsorship and resource support.• Successfully planned and executed digital marketing strategies, reaching over 50,000 individuals.