# Hugh Nolan:Electronics Engineer (Biomedical)

Hugh Nolan

Electronics Engineer

I work at The Irish Longitudinal Study on Ageing (TILDA, www.tilda.ie), which is a 10 year study of 8500 people over the age of 50. We aim of collect data covering many aspects of older people’s lives, from health to family and friends to economic circumstances, in order to determine important factors in healthy ageing, improve quality of life and investigate causes of various health conditions. I am responsible for analysis and translation of raw collected data from a wide range of systems into meaningful outcomes which can be incorporated into our analysis dataset and understood by non-technical researchers. Some of the varied data types that I work with are: electrocardiogram, blood pressure + heart rate, near-infrared spectroscopy (brain blood flow measurement), respiration and 7-day accelerometer recordings, and will soon be working with functional MRI. I am also responsible for equipment maintenance, IT planning and support, and data security.

How do you use mathematics within your job?

I use mathematics to analyse and transform data, detect patterns and remove bad data, and create summary measures from data files of 10,000 – 100,000,000 data points. Since we have such recordings for up to 8,500 people, I also must automate these analyses, creating robust algorithms that work well with measurements from people with various health conditions which affect the data, along with coping with bad records (loose connections, etc).

What type of mathematics do you use to solve problems?

I use a wide range of mathematics, depending on the type of data/problem I am looking at. Some of the common methods and approaches I use are:
Time-series analysis: visualisations, data filtering, smoothing, Fourier / wavelet analysis, peak detection, warping, segmentation
Matrix manipulations: averaging and combinations, normalization, noisy data detection, principal component analysis, independent component analysis.
Statistics/probability theory: change detection, outcome comparison, data prediction, mathematical modelling+data fitting, identification of important time-points
Transformations: logarithms, exponentials, power transforms, statistical transformations

What aspects of the mathematics curriculum or mathematics courses have proven most useful to you?

Calculus and linear algebra underlie a lot of the methods I use to transform data – I do not (usually) need to manually solve equations to use these methods, but I do need to be aware of the principals of how they work to know when is appropriate to use them. Geometry and trigonometry were very important for developing a sense of intuition about solving many real-world problems – many data analytics problems can be viewed geometrically. The circle and the sin and cos functions are fundamental concepts in signal analysis. Basic set theory is very useful for manipulation and alignment of groups of data. Statistics is incredibly useful in research and computing fields; unfortunately the Leaving Cert statistics/probability courses were not sufficiently detailed/practical to give me a useful background in the field.

What is your education to date?

Bachelor of Electronic Engineering degree at UCD.
Postgraduate Diploma in Statistics at TCD.
PhD Neural Engineering in TCD.