We present PyTEDA-Web , an interactive web platform for running reproducible data assimilation (DA) benchmarks with real-time progress streaming and persistent experiment tracking. PyTEDA-Web is built on top of the existing PyTEDA scientific package Niño-Ruiz (2025) 1, and focuses on orchestration: it provides a FastAPI backend to create and manage DA runs, a PostgreSQL persistence layer to store requests, method-level metrics, time-series outputs, and event logs, and a lightweight browser-based frontend that enables interactive comparisons across multiple ensemble Kalman filter (EnKF) variants. Progress and intermediate results are streamed to clients in real time via Server-Sent Events (SSE), enabling users to monitor long-running experiments without polling-heavy workflows. PyTEDA-Web targets educational and research use cases where rapid experimentation, method benchmarking, and transparent run provenance are essential. The platform is available as an online demo as well as a self-hosted application that can be deployed locally or via Docker.
Elías D. Niño-Ruiz (Sat,) studied this question.