Proceedings of the International scientific and practical conference ―Oxford International Science Forum‖ (April 17-19, 2026) / Publisher website: www.naukainfo.com. - Oxford, United Kingdom, 2026. - 367 p.

103 Explanatory signals support model interpretation, such as through feature importance or local explanatory ratings [4-6]. Major crowdfunding platforms and their analytical limitations . The analytical characteristics of CD strongly depend on the platforms from which the data originates. Among the most representative examples are Kickstarter, Indiegogo, GoFundMe, and Wefunder [7-10]. Kickstarter is one of the most well-known rewards-based platforms, and it is analytically attractive due to its all-or-nothing funding logic, which provides a relatively clear interpretation of success and failure [7-8]. However, this same clarity has its limitations: the platform does not function as a fully open research data environment, and detailed information about sponsors is available mainly through author-oriented reports, rather than through public access to large-scale studies [7-8]. Indiegogo is another important rewards-based platform that offers creator- centric workflows and public API resources [9]. However, the availability of developer access does not eliminate methodological challenges: changing platform architecture, evolving campaign logic, and historical inconsistency can reduce comparability over time and complicate longitudinal studies [9]. GoFundMe represents a donation-based model where narrative credibility, urgency, and emotional resonance are central to campaign effectiveness [10]. This is important for data analysis because many of the influential features in such campaigns are behavioral and textual rather than reward-related. A limitation is that broad public access to the platform‘s detailed, research-oriented data is limited, while API usage is mostly limited to professional-grade integrations [10]. Wefunder reflects an equity-based model and therefore contains stronger semantics related to investment and community investment [11]. This increases the value of such data for financial and venture analytics, but makes direct comparisons with reward-based or donation-based platforms difficult, as the operational meaning of ―success‖ is different [11]. Thus, the key conclusion of the review is that the major crowdfunding platforms are not analytically equivalent. Their differences concern not only openness and