Freelance data science work
Review of study design, suggested analysis approach, exploratory data analysis requirements.
Proficient in Frequentist Statistics, Machine Learning, and Nonlinear Time-Series Analysis.
Available to present on various topics in data science (topics available upon request).
Tasked with researching and creating bespoke analysis methods to monitor, predict
and forecast changes in Multiple Sclerosis (MS) based on data derived from smartphones. Inspiration for these
approaches has come from complex systems science (e.g.
Recurrence Quantification Analysis),
optimal control theory (e.g.
Licitra et al., 2018),
electrical engineering (e.g.
Single Input Single Output systems),
and neuroscience (e.g.
Subtype and Stage Inference).
Company -
Neurocast
Duration - 1yr [2021 to present]
Location - Amsterdam, Netherlands
The main data scientist assigned to conduct analyses in collaboration with the Amsterdam University
medical center. This collaboration, intended to validate the use of smartphone interactions as a means to
monitor and predict Multiple Sclerosis (MS) outcomes, led to
a cross sectional journal article,
a responsiveness journal article, and
a longitudinal journal article.
Company -
Neurocast
Duration - 1yr [2020 to 2021]
Location - Amsterdam, Netherlands
Created a
machine learning workflow
and conducted an assortment of machine learning
analyses (SVM, Random Forests, Multilayer Perceptron etc.) to predict MS outcome
measures using keystroke data.
These analyses were all internal but an example of an approach similar to what was
done may be found here.
Company -
Neurocast
Duration - 1yr [2019 to 2020]
Location - Amsterdam, Netherlands
Applied nonlinear time series analyses - specifically dynamic complexity to keystroke dynamics
in order to match fluctuations in the keystroke data with clinically relevant changes in outcome
measures. This culminated in the following journal article: Twose
et al, 2020. Also looked into the use of Mixed Effects Random Forests to predict
MS outcomes whilst taking into account the hierarchical nature of clinical data and still
applying a nonlinear approach (in contrast to Linear Mixed Models).
Company -
Neurocast
Duration - 1yr [2018 to 2019]
Location - Amsterdam, Netherlands
Created a functioning medicines relational database;
researched vector autoregressions to predict adverse events;
researched scientific advices to change the protocol the CBG
uses to regulate antidepressants.
Company -
CBG-MEB
Duration - 1yr [2015 to 2016]
Location - Utrecht, Netherlands
Measuring perceived stress levels, coping strategies and
physiological correlates in UCR students at times of high and
low workload: Selected to undertake this study for the
management of University College Roosevelt to influence policy
decisions on the protocol the university uses to combat stress
in students.
Company -
University College Roosevelt
Duration - 1yr [2015 to 2016]
Location - Middelburg, Netherlands
