Projects.

Research is ups and downs;

​

late nights and early mornings;

​

some success and many failures;

​

and most of all;

​

reading, reading, and more reading. 

Scroll Down For More.

What Factors Contribute to "Successful" Spotify Collaborations?

DS4A Empowerment | Project | 2021 

• ​Led team to develop a clean project employing 3 years of data collected from Spotify Top 200 Charts in order to examine factors affecting the success of an artist collaboration. 

• Worked to develop the application wireframes, checked and improved Python code throughout the entire project, finalized the exploratory data analyses, and built the Dash application. 

• Project was recognized for its clean presentation, interactivity of the dash application, and the breadth of work completed.

• Team GitHub Link: https://github.com/PeterGarcia95/spotify_collabs

• My GitHub Link: https://github.com/JPSchloss/Spotify-Dash-Final

• Interactive Application: https://spotify-collaborations-app.herokuapp.com/ 

Carolina Data Challenge Text Processing Session

Carolina Data Challenge | Bootcamp | 2020 

• Taught introductory text-processing course to about 30 participants as part of the Carolina Data Challenge. 

• Created a simplified lesson around the actual process of text processing and cleaning, simple sentiment analyses, and topic modeling using R and RMarkdown. 

• Told that it was one of the best sessions throughout the event by participants and the event liaison. 

• GitHub Link: https://github.com/JPSchloss/CDCTextProcessing

"Bleach, News, and Tweets” | 3MT Presentation

Three-Minute Thesis Competition | Presentation | 2020 

• ​Created an engaging and informative presentation that communicated a research project in under three minutes.

• This is part of the Three-Minute Thesis competition that is hosted by universities worldwide.

• Was chosen as a finalist for my university and was able to present my research live.

Alt-Right Rhetoric and Conspiracy Theories:

Topic Modeling History Channel’s ‘Ancient Aliens’

UNC Chapel Hill | Independent Study | 2021 (In – Progress)

• Led research examining how alt-right rhetoric promoters use conspiracy spaces to introduce extremist ideologies into mainstream society, how that rhetoric is presented, and whether it has grown over time. 

• Web-scraped, cleaned, stemmed, employed three different topic modeling approaches (STM, VEM, and Gibbs), and ultimately deployed an RShiny application utilizing 180+ television show scripts from the first 15 seasons of History Channel’s ‘Ancient Aliens’. 

• This work found alt-right themes appearing within the show’s topics, growth in the presence of alt-right rhetoric over time, and the incorporation of a greater diversity of conspiracy theories over time. 

• Interactive Application: https://jonathanschlosser.shinyapps.io/AncientAliensApp/

Coronavirus and Conspiracy Theories:

What are the news media and public discussing? 

UNC Chapel Hill | Course Project | 2020

• Worked to collect 3000+ news articles and 250,000+ tweets using coronavirus and conspiracy theory related search terms in order to identify the similarity between the public conversational sphere and the news.

• Employed topic modeling, natural language processing, and some light machine learning to identify topics present in the two conversational realms and to identify their similarities. 

• Found that the overall topics were similar in the two realms, but that the specific language used was different, showing a disconnect in how the two realms actually discuss things. 

• GitHub: https://github.com/JPSchloss/CovidConspiracies

• Interactive Application: https://jonathanschlosser.shinyapps.io/COVIDConspiraciesApp/

The Disinformation Ecosystem and the Frames Employed

UNC Chapel Hill | Course Project | 2019

• Examined the topics and frames employed in 2.9 million Tweets originating from Russian disinformation actors during the 2016 US election through text mining and topic modeling approaches.

• Worked to identify oppositional topic framing being presented by disinformation actors; including the simultaneous dissemination of both Black Lives Matter and Blue Lives Matter content, as well as pro-Clinton and pro-Trump content.  

• Concluded that disinformation actors were presenting frames on both sides of an issue to “muddy the waters” and sow discord in US political discourse. 

Disinformation Tweets and News Articles – Exploring The Discussion Similarities with Topic Modeling.  

Summer Institute For Computational Social Science | Group Project | 2019

• Employed topic modeling and NLP procedures to examine two datasets, including a 2009-2018 dataset of IRA tweets and a 2012-2016 dataset of Facebook posts from the top 15 news organizations in the US, to identify similarities in the language being used between disinformation actors and news actors on social media.

• Found that there were only small similarities in the topics being discussed, but larger differences in the rate of discussion around the topics; such that disinformation actors talked about a narrow set of topic areas compared to news outlets, while discussing those topics to a greater extent. 

• Led the team to deliver the project and presentation, and specifically worked on data collection, data cleaning, assisting in the topic model build, and creating data visualizations. 

Social Media, Environmental Twitter, and Diverse Viewpoints

Lancaster University | MSc Dissertation | 2016

• Conducted a basic text-mining analysis of 60,000+ Tweets mentioning the 2015 Paris Agreement focused on identifying the presence of influential actors and frequency of terms, the evolution of retweets throughout the conference, and the uniqueness of the conversation. 

• Employed a novel content-analysis framework (based on Norman Fairclough’s Critical Discourse Model) to review 60,000+ tweets and identify the most ‘powerful’ tweets, users, and tags in the conversation.

• Found that the conversation centered around a few very similar tweets from highly influential actors (e.g. UN, President Obama, and Prime Minister Modi) that were retweeted widely (contributed to more than 70% of the sampled discussion).

Social Media: A Source of Geospatial Information

Lancaster University | Project | 2015 

• Conducted an in-depth literature review on the use of social media as a source of geospatial and contextual information for GIS queries.

• Analyzed the risks and benefits including the possible invasions of privacy, the limitations of social media- sourced GIS information, and the related costs, time investments, and overall efficiency.

Costa Rican Ecosystem Survey

Binghamton University | Project | 2014 

• ​Traveled throughout Costa Rica for ten days to explore the variety of ecosystems in the country including the Tropical Rain Forest, Cloud Forest, Paramo, and Mangroves.

• Surveyed the differing landscapes through plant/animal identification, photography, and note-taking, in order to gain an in-depth understanding of the key characteristics.

Climate Action Survey

Sullivan County Planning Board | Internship | 2013 

• ​Researched and prepared greenhouse gas emission surveys for two municipalities and the Sullivan County Community College as part of the 2014 Sullivan County Climate Action Plan.

• Performed Scope 1 and Scope 2 analysis on municipality owned buildings, motor vehicle fleets, and overall energy use in order to identify prominent emission sources and to provide accurate recommendations for mitigation.

• Work has been used across Sullivan County to guide sustainable building renovation, vehicle fleet upgrades & renewable energy decisions.

Drop Me A Line.

I'd love to hear from you.

Feel free to follow me on social media.