Microsoft Azure Industry Project

I worked in a team of 4 Master's students to create user-friendly end-to-end data science notebooks for Microsoft Azure.

Our Github

Parking Lot Utilization Study

Collaborating with the Town of Amherst, I worked in a team of 3 graduate students and two professors to develop a deep learning model that predicted parking lot occupancy for the Town of Amherst. We used data science and software engineering techniques to answer the pressing question in Amherst: how bad is parking availability?

Our Github

My D3 Visualization

Audie the Bear

I participated in SheHacks, the world’s largest all-female and non-binary hackathon in January 2018. In 36 hours, a team of three hackers and I created Audie. Audie won the following awards, Competing with over 150 teams of hackers:

1. She </SMILES>
2. Best Hack that Makes the Process of a Hospital Visit Better

Purpose

Audie is a bear made to make improve hospital experiences for children. Audie's purpose is to facillitate discussions with children about feelings, especially for children who may experience difficulty expressing their feelings directly to adults.

How Audie Works

Audie's goal is for patients to feel heard, but not intimidated by their patient experience. Audie is a bear that is activated by sound and records speech. Patients will talk about how their feelings and Audie will encode their words into a sentiment value once a recording has been made. This value reflects how positive or negative what the patient says is. From here, a doctor can access the sentiment data of the patient and make an assessment on the patient's progress. To do this, we leveraged Google Cloud's NLP and Speech APIs, MATLAB, and an Arduino circuit.

Learn more about Audie at:

Audie's Github

Audie's Devpost

Consulting Website

Purpose

I created a professional website for Jennifer Stine's private consulting. This site is hosted at: www.gobrighter.net

Method

This was my first experience in creating a website from scratch. I decided to use three.js to add interactive web graphics to enhance user exprience. Additionally, fullpage.js and Bootstrap were used for formatting and styling.

Visualizing Movie Data

Big Data Analytics Final Project

In this final project, I analyzed movie data with Apache Spark data frames and created interactive visualizations with Apache Zeppelin. The dataset used was provided by MovieLens and contains over 45,000 movies. Data were stored in the ml-latest.zip file which contains 224 MB of data.

Purpose

Utilize Zeppelin and Spark Technologies, as well as SQL queries to create interactive and dynamic visualizations of the MovieLens dataset and discover how the movie landscape has changed over time.

Notebook

The Zeppelin Notebook is stored in two links:

• Notebook
• Source Code

Example from Notebook:

Personalized Multimodal Urban Travel Planning

Senior Capstone Project

For my senior capstone project, I worked in a team of students, an academic advisor, and a Ford data scientist sponsor, to create a multi-modal trip planning tool for the Boston metropolitan area that recommends personalized travel. This project won the "Biggest Impact" Capstone Award, competing against over 20 capstone teams.

Problem Statement

Transportation is constantly growing and evolving to meet the demands of people. Modern travelers, especially in urban settings, have a wide variety of transportation options. Therefore, this project is focused on building a multi-modal trip planning tool for the Boston metropolitan area as a case study. The tool will recommend personalized travel, based on a user’s preferences surrounding travel time, delays, costs, number of transfers, and comfort.

Deliverables

• Exectutive Summary
• Final Presentation
• Final Report
• Trip Planning Tool
• Poster

Individual Contributions

• Organized the case study data analysis and designed the case study diagrams
• Created data visualizations for the data gathered from MBTA and Hubway with Tableau
• Collaborated in mathematical model selection and formulation
• Completed "Data Collection and Analysis" section of the Report and Poster
• Built the algorithm and user interface in Excel

Credit Suisse Capital Allocation Paper

During my co-op at Credit Suisse HOLT, I had the privilege of working with the Corporate Insights Team to prepare data for the Capital Allocation Paper. Collaborating with the Corporate Advisory team, I leveraged HOLT's Oracle database of over 20,000 companies’ financial information to visualize data for a capital allocation publication that was shared with CEOs of major corporations.

Publication

The paper was published and distributed by Credit Suisse. The online publication can be found here.

Individual Contributions

• Collaborated with the Corporate Insights Team on data exploration
• Queried data in SQL to classify companies as "Returners" or "Reinvestors" (See Fig. 2)
• Visualized the data with dynamic spreadsheets to aid research
• Recognized for data analysis (See Page 15)