Data Science and Engineering Portfolio
This page showcases some of the projects I have worked on related to the field of Data Science and Engineering. Data Science and Engineering is a field I find particularly invigorating.
IBM Data Science Capstone Project (Coursera)
The final project of the IBM Data Science course offered on Coursera is to predict if the first stage of the SpaceX Falcon 9 rocket will land successfully.
In the project we collect and wrangle the data. We explore the data by visualizing it and using SQL commands. We create an interactive Folium map and a dashboard. Finally we train and evaluate 4 classification models using logitistic regression, SVM, classification tree and k nearest neighbors.
Relevant skills: Jupyter Notebook, RESTful API, Beautiful Soup, Seaborn, Matplotlib, Plotly, Folium, Dash, SKLearn, panda, numpy
Flight Simulator
The initial goals were to learn more about the Godot Game Engines and to create a simple project which would help online students learn about modelling and simulation. The course unfortunately fell was put on the back burner, but I did learn a lot about Godot.
With a bachelor degree in Aerospace Engineering simulating the dynamics of the aircraft was a fun challenge. Furthermore I learned a lot about the Godot Game Engine.
Relevant skills: Modelling and Simulation, 3D graphics
Airbag Calibration Toolbox
Not only does the airbag algorithm needs to recognize a crashes within the blink of an eye, but it also needs to be stable enough to avoid deploy when the accident is not severe. While working as an airbag application engineer for the company Robert Bosch part of my responsibility was to tune the parameters of the complex airbag deployment algorithm to fit the specifics of the client's model.
In order to assist the calibration process I developed a toolbox in Matlab. The toolbox included multiple functions to visualize crash data and to investigate the effect of changing different parameters. This made it easy to iterate across multiple parameter values and pick the optimal one. Furthermore I implemented a simple optimization algorithm which would pick a set of parameters at the local optimum.
Relevant skills: Matlab, Visualization, Optimization
Model Predictive Control of a Wave Energy
Converter
The Archimedes Wave Swing (AWS) is a wave-energy converter that generates electricity from ocean waves. The control challenge is to maximize the energy produced while satisfying the constraints on the motion. The control problem is made more challenging because ocean waves are highly irregular.
For my masters thesis at the TU Delft, I developed and evaluated a model predictive controller. The controller uses an ARS model to continuously predict the excitation force due to the incoming wave and calculates the constrained optimal control force to safely maximize power production.
Relevant skills: Matlab, Modelling and Simulation, Identification, Model Predictive Control, Optimization