An unconventional route to data

My route into data analysis was not a straight line: a degree in Anthropology, a Master's in Forensic Sciences, and now one in Data Analysis and Decision Support Systems. Looked at carelessly, that is a set of jumps between unrelated fields. But one question runs underneath all of it: how a structure keeps holding information about something that is no longer there.

I have spent time on both sides of that problem. In the shell middens at Muge I dug for fragments that meant very little on their own and a great deal once you knew what they had been lying next to. In a Napoleonic-era infirmary at the Convento de Santa Clara in Coimbra it was the same: the individual finds mattered far less than the arrangement they were found in. My first thesis estimated age at death from a single bone, the clavicle — not because a clavicle records an age, but because the way its parts change together narrows the range of ages that remain possible.

A missing value in a dataset is that same problem on different objects. The value disappears as an observation, but not as an element in a structure of dependencies: while the relationships hold, so does information about what is absent.

Business analysis turned out to be the same exercise again. Every analysis is a search for something that was not visible before it started — and the part I like most is going after what was hidden well enough that nobody thought to ask the question in the first place.

I like doing that work myself. I would rather spend an afternoon writing the code and sketching the analysis than press a button and check the output at the end of the day. Most of what I come to understand about a problem arrives during that afternoon, not after it.

Currently in Coimbra, working mostly in Python, SQL and Power BI, with a particular interest in statistical methods and Bayesian approaches. Always up for a problem where the data turns out to be worse than it looked.

Education