Consumer Militancy
Description
This repository contains the dataset and analysis code used in the manuscript "Consumer Militancy: How Entrepreneurial Desire Influences Technology Adoption," which examines the influence of entrepreneurial desire on consumer behavior in technology adoption. The dataset includes a SQLite database containing comments posted on videos published in an YouTube channel and a spreadsheet with text fragments extracted from a keynote speech. The analysis code, written in Python, is provided to ensure transparency and reproducibility of the research findings.
Files
Steps to reproduce
To reproduce the results presented in the manuscript "Consumer Militancy: How Entrepreneurial Desire Influences Technology Adoption," follow these steps: Data Preparation: Ensure that the files provided in this repository are placed in the same directory. The scripts assume that the data is in the same directory as the code. Dependencies: The analysis was conducted using Python with the following libraries/packages: pandas, numpy, matplotlib, umap-learn, scikit-learn, sentence-transformers, torch, tqdm, scipy, seaborn, pingouin, scikit-posthocs. Install these packages and their dependencies with 'pip install -r requirements.txt' Running the Scripts: The analysis is divided into multiple scripts. To reproduce the results, run the scripts in the following order: 1. create-embeddings-STEVE-JOBS.py: Run this script first to create a vector embedding of each text fragment in the spreadsheet. 2. create-embeddings-APPLE.py: Next, run this script to first select the Apple sample from the SQLite database and then create a vector embedding for each comment. 3. create-embeddings-SAMSUNG.py: This script selects the Samsung sample from the SQLite database and creates the vector embeddings. 4. calculate-similarities-comments.py: Computes the semantic similarity between the entrepreneur's revolutionary discourse centroid and the comments from each sample. 5. analyze-similarities.py: Compares the two distributions of semantic similarities. 6. umap-plot-centroid-APPLE-SAMSUNG.py: This scripts creates a 2D plot containing the centroid and 10,000 embeddings of each sample Output: The results, including figures and tables, will be saved in the directory. These outputs should match those presented in the manuscript. Customization: The code is well-documented, allowing for easy modification and adaptation to different datasets or parameters if further analysis is desired.