Data Management Plan — (28 Apr 2025) for the NFL-Attendance-RF projectSubmitted for Part 2 “Data Management” — Data Stewardship UE 2025S, TU Wien 📂 Input & derived data – Kaggle source tables and the four curated splits (train, validation, test, merged baseline) are published in DBRepo (DOIs P1–P4). 🤖 Model & results – The final Random-Forest regressor plus five evaluation/diagnostic artefacts live in TUWRD(DOIs O1–O6). 🧩 Rich metadata – A FAIR4ML record is embedded in the model landing page, and an external CodeMeta file connects author, dependencies and every PID. 🔒 Preservation & security – Redundant storage (GitHub → Zenodo, DBRepo, TUWRD) and TU Wien’s ten-year retention guarantee long-term accessibility. 💻 Full code — notebook, helper scripts, requirements.txt, README, licence and ... — is on GitHub:https://github.com/emilp-tuwien/nfl-attendance-prediction (DOI: https://doi.org/10.5281/zenodo.15292895; clone the repo to rerun the entire pipeline).