Welcome to the TikTok Tracker project repository! This initiative is part of Dr. Emily Hand's CS 791 Interdisciplinary Data Science Course for Spring 2024. Team members Lauren Feldman, Joey Paschke, and Ryan Baldwin dive deep into TikTok video data to uncover patterns that could predict trending content.
TikTok serves as a crucial platform for the spread of information. In this project, we focus on understanding how hashtags may influence the virality of posts. By analyzing data from 1,867 unique TikTok videos, including metrics like like counts, follower counts, timestamps, and hashtag usage, we strive to determine the impact of these elements on content popularity.
Through our investigation using k-nearest neighbors clustering (KNN) and a decision tree classifier (DT), we discovered that hashtag usage does not significantly influence the popularity or reach of a post. This README covers our approach, findings, and discusses ways to refine our prediction models.
- Analysis of 1,867 TikTok videos
- Utilization of KNN and DT algorithms for trend prediction
- Exploration of hashtag impact on video popularity
We employed two main predictive models:
- K-Nearest Neighbors (KNN): For clustering based on similarity measures.
- Decision Tree Classifier (DT): To classify data into categories based on learned decision rules.
The effectiveness of these models was assessed through various metrics, which are detailed within our project documentation.
To get started with this project:
- Clone this repository: Ensure you have Git installed on your system and run the following command:
git clone https://github.com/your-username/tiktok-tracker.git
- Open in Google Co-Lab: Download the project files from this repository and upload them to Google Co-Lab for an optimized viewing and interaction experience.