Kickstarter crowdfunding dataset for machine-learning and sentiment-based prediction of project success (2009–2026)

Résumé fourni par la source

Processed Kickstarter dataset supporting the study "Machine learning and sentiment analysis for intelligent project selection: design and experimental validation of a decision support workflow" (Kobets & Starchenko, Herald of Advanced Information Technology, 2026). The data derive from the public Web Robots Kickstarter export (83 monthly CSV files, snapshot 12 January 2026), merged and deduplicated to 204,907 unique projects (2009–2026). It includes the cleaned master table, the model-ready feature matrices (40 structured features; with and without VADER/FinBERT sentiment features), project-level comment-sentiment aggregates for the 2,085 projects that overlap the comment corpus, and the result tables behind the paper. The labeled set contains 194,388 projects (59.8% successful / 40.2% failed). See README.txt for full variable descriptions and the mapping to the paper’s tables and figures. Original Kickstarter data © Web Robots (webrobots.io).

Ce résumé expose les affirmations des auteurs. BNTIC ne l’interprète pas comme une validation indépendante des résultats.