NFT Bidding Behaviours and Sentiment Analysis
Description
This folder serves as a repository for hosting the raw data and jupyter notebook files needed to recreate the analysis for 'Dutch auction dynamics in non-fungible token (NFT) markets', published in the Economic Modelling journal. Purpose of File 'Telegram_Sentiment_Analysis_using_VADER.ipynb': 1. Deconstructing and cleaning 30,197 Telegram Forum messages to extract further insights on the bidding and listing behaviours of investors involved in a fledgling NFT marketplace 2. Applying a Sentiment Analysis model specialising in social media communications (VADER) 3. Extracting sentiment scores, for later analysis against NFT listing and sales data 4. Visually inspecting the evolution of forum participation and investor sentiment over an 8-week study period. Purpose of File 'NFT_Plots.ipynb': 1. Visually inspecting the evolution of 28,919 NFT listings and 4,937 NFT sales data over an 8-week period. The data required to run both above notebooks are available in the 'NFT_Dutch_Auction_Study_Data_Regressions.xlsx' file Purpose of File 'NFT_Dutch_Auction_Study_Network_Graph_of_Wallets.ipynb': 1. Visually inspect the inter-connections between all wallets buying and selling on the NFT marketplace, where each wallet represents a node. This is one method of revealing suspicious trading patters amongst marketplace users. The data required to run the above notebook is available in the NFT_Dutch_Auction_Study_Data_Network_Analysis.xlsx' file Jupyter Notebook(s) Author: Dr. Darren Shannon. All errors my own. Journal Article Authors: Dr. Darren Shannon, Prof. Michael Dowling, Dr. Marjan Zhaf, Dr. Barry Sheehan More information on running the files: - Minor code amendments are required in each notebook to open the data file and save generated images locally. - Although minimial code adjustments are required to get the code working, it is suggested that you have a beginning level of proficiency in Python to run the code. - Virtual environment / project generation / library / root files are not provided and will need to be created before running the Python code. - Information on the library packages required to run the analysis are provided in the relevant notebooks. - Please direct any detected errors in the code / data to: darren.shannon@ul.ie