Introduction
Fernando is a data scientist who leverages large datasets and multiple data sources to gain operational insight on public transport systems in London, aiding the build of predictive statistical models that result in improved network reliability.
This Portfolio aims to show both personal and professional projects (when the code can be shared openly).
Deployment of ML model to detect credit card fraud as microservice
This two part post showcases how R can be used on a professional enviroment to generate production level results.
Part 1 details how to set up a quick classifier model with extreme gradient boosting and mlr.
Part 2 shows how, with the help of a RESTful API using plumber and docker, we can containeraize our application to be deployed in the cloud to quickly integrate it within our company.
bikecommclust
Complementary code for “Community structures, interactions and dynamics in London’s bicycle sharing network”
Oct 8, 2018, Ubicom PURBA-2018
We apply a novel clustering technique to London’s bikesharing network, deriving distinctive behavioral patterns and assessing community interactions and spatio-temporal dynamics. The analyses reveal selfcontained, interconnected and hybrid clusters that mimic London’s physical structure. Exploring changes over time, we find geographically isolated and specialized communities to be relatively consistent, while the remaining system exhibits volatility. We increase understanding of the collective behavior of the bikesharing users.
Blog Post - Medium @ Towards Data Science
saturnR
The package saturnR is a thin wrapper around some complex system calls that automate highway model builds as well as data extraction for output processing on the SATURN software. Github page (Private)
spellbook
Collection of misc functions used across different projects as an R package facilitating reusing bespoke code on transport analysis and time series tasks.
Simplex Optimisation
Companion code for essay: “Optimum Design of Prestressed Beams under Uncertainties in the Tendon Layout” (2015) While not exactly reproducible showcases an implementation of linear optimisation algorithms on FORTRAN with multiple constrains. While the techique is admidetly not the best suited for such a non-liear problem it’s a great example on linear algebra algorithmic implementations.
Fernando Munoz 2019