Back to projects
Data / AI
CFMDG Dashboard
A data science Dash app around the CFMDG dataset, with .dat parsing, exploration, regressions, Ridge/Lasso models, Gradient Boosting and Gaussian Process models.
Context
Data/AI project focused on predicting Smax from coastal scenarios and time series.
Role
Handled data loading, feature extraction, Dash tabs, model training and metric analysis.
Stack
PythonDash
Architecture
Python Dash app with data_loader, lru_cache, modular tabs, Jupyter notebooks, shell scripts and scikit-learn models.
Key features
- Exploration of NM, T, S, WaterLevel, Hs, Tp, Dp, U and DU variables.
- Multiple regression, logistic regression, Ridge/Lasso, Gradient Boosting and Gaussian Process.
- R2, MAE, RMSE, F1 and ROC metrics depending on available models.
Learnings
- Transform scientific files into usable features.
- Compare linear, ensemble and Gaussian Process models.
- Present data results in a readable dashboard.
Gallery