Galvanize Capstone ProjectDocument Classification Stack Overflow Tag Classification (NLP) Project Link: https://github.com/Denton044/StackExchange-Analysis● Problem: Document classification is used to divide and organize text for fast, efficient retrieval in the future. This project explores various NLP techniques to automate this classification task on a dataset from Stack Overflow.● Objective: Stack Overflow is a knowledge sharing platform where users may post questions and answers. Each question is manually tagged with categories by the user, known as “tags”. Build models that can correctly predict the appropriate tag(s) for a given question based solely on the text in the title and body of a given post.● Data: Kaggle – The dataset was obtained from Kaggle. This analysis will look at the top 15 most frequently used tags in the dataset (~1M questions total).● Tools: TF-IDF, Word2Vec Skipgram, Random Forest, Logistic Regression, SVC, Multinomial Naive Bayes, CNN.● Results: 4 traditional algorithms were considered with a TFIDF vectorizer. Multinomial Naive Bayes, Random Forest, Logistic Regression and Linear SVC. I also used a 2 layer 1 dimensional CNN with a fully connected ANN output layer in conjunction with a Word2Vec Skipgram model. The CNN model outperformed all other models with a F1 score of 0.77.

• Analyze API metrics with client usage data. Assess API performance and track client use patterns.• Develop and implement API daily data regression monitoring tests to ensure high degree of data quality and consistency via Python scripts for Xignite’s new APIs.• Assess client specific use case to determine how they can leverage Xignite APIs to achieve goals. Provide support on Xignite’s catalog of hundreds of APIs. Develop hypothesis of underlying root cause of bugs and work with engineers to restore expected behavior.

• Review and maintain client balance sheets; analyze cash flows, portfolio performance, and asset allocation for over 35 high net worth individuals• Direct and manage workflow for team’s junior analyst• Provide robust and detailed financial models to facilitate executive level decisions• Develop comprehensive client-facing reports, abstracts, and charts to summarize and guide investment strategy for existing and prospective clients • Execute trade orders and facilitate money transfer requests for clients on a daily basis

• Created and assembled customized financial reports outlining historical performance, asset allocation, and portfolio comparisons for individual clients • Provided retirement planning data analysis via Monte Carlo Simulation• Evaluated fund managers to assess suitability for clients based on risk, return, and market trends• Compiled and oversaw database of key client information such as net worth, assets under management, meeting dates, and issues to be addressed for subsequent formal meetings

• Designed and executed experimental procedures for testing the uptake of biochemical compounds critical to the folate transport system in cancer cells • Analyzed the cellular uptake of compounds critical to the folate transport system in cancer cells using fluorescence microscopy and statistical methods