Development of a Predictive Model Deployment and API Platform

Data Science — 2025, Undergraduate

This study developed a predictive model deployment and API platform that lets data scientists register trained models, supply test inputs, and serve predictions over a REST API with usage logging. The platform compares model versions. Adopting a client-server architecture and a survey-based usability evaluation, the system was built using Python, Flask, and SQLite. The findings showed that the platform simplified model serving and accelerated integration, while achieving a high System Usability Scale score. The study recommends use in data science teams.

Homepage screenshot

Report a problem with this project