marisapapagelis Goto Github PK

6.034-mit-artificial-intelligence

Labs for 6.034 Artificial Intelligence Course at MIT

car-value

Supplementary material for the working paper entitled "The Value of Car Ownership and Use in the United States"

company-data-analysis-tgh

cs304-project

DoorToDoor is a platform to help students open doors throughout the challenging recruitment process. Our platform will help students in all stages of the process, from first-years who want a better understanding of different industries to seniors who need assistance with interview preparation and networking with alumni. DoorToDoor will connects students with alums in various industries, allows students to share resources, interview questions, experiences with employers, deadlines, and provides anything else a student might need. There are many existing platforms, but we want to bring it all together in one and solely focus on the Wellesley community.

cs307-computer-graphics

data-science-capstone

Wellesley College Data Science Major Capstone

facial-recognition-algorithms

hobbyhacks-tiktok

HobbyHacks Hackathon by TechTogether

media-partisanship-2020

Facebook is becoming an increasingly popular distributor of political news, especially among younger generations. As the 2020 U.S. presidential election approaches, news outlets are reporting on a variety of political topics, and their readers are interacting with them on a widespread scale. This project uses these ideas to conduct an investigation into the most frequently discussed political topics across partisan and nonpartisan news sources on Facebook and builds a classification system to predict partisanship of news posts. To accomplish this, the code for the project uses Selenium with Chromedriver and BeautifulSoup to scrape, store, and parse publicly accessible posts on Facebook, and the NLTK package for Python to create a supervised classification system.

multivariate-data-analysis

portfolio

regression-analysis-boston-marathon

stat228-multivariate-data-analysis

technologies

teen-vape-classifier

The purpose of our project is to analyzes the factors contributing to teen vape use. Using data from the 2014-2018 New York Youth Tobacco Surveys, a classification tree was trained with eight factors to predict teen vape behavior and determine which contributed most to vaping. Based on the analysis, the biggest contributing factors were having lived with a smoker, whether the student is in high school or middle school, and allowance money. Validation of the model yielded an accuracy of 74.13%.

twitter-bots-election

Leading up to the 2016 United States presidential election, the Russian government infiltrated the United States with thousands of fake social media accounts with the intent of spreading misinformation and harming Senator Hillary Clinton’s presidential campaign. False information from Russian government-controlled media reached millions of American social media users between 2013 and 2017. TwitterTrails, a Web-based investigative tool created as a research project at the Socal Informatics Lab at Wellesley College, analyzes the origin and propagation of rumors from accounts on twitter. In this project, I investigate data on Russian Twitter accounts to create, analyze, and depict a visualization of account users and stories. The data used in this project is stored in file containing the screen name, user identification number, tweet count and story count or each Russian account. Lastly, the data for each account includes the identification numbers of all of the stories that the account interreacted with. Both the data and the visualization will be used in the final part of the project to draw conclusions pertaining to the use RATs leading up to and during the 2016 Presidential Election.

wdc-honors-theses

a web scraping tutorial and website for Wellesley College honors theses archives as part of the Wellesley Data Collective project

yelp-nlp-profile

The purpose of our project is to develop a web application that scrapes consumer reviews on Yelp to build profiles that showcase their knowledge and expertise. Additionally, the research will utilize natural language processing to analyze data and build user profiles. As a result, Yelp users will be able to share their profiles over social media, which will further incentivize them to continue reviewing and contributing quality content.