NEXT CANADIAN FEDERAL ELECTION --- days
I do data analysis.
Political, electoral, economic, financial.
We're no good at making sense of lists of numbers. Data science exists not just to create human-readable narratives about our data but to infer new information. Data also needs to be understood in its own context. While there is some truth to the idea that "it's all just numbers", the interpretation of those numbers and the relative usefulness of any conclusions drawn rely heavily on understanding where they came from, how they interact in situ, and, importantly, which ones are missing.
Areas of Interest
The link between thermomdynamic entropy and information entropy established by Shannon in 1948 is fascinatingly intuitive. Landauer went on to put a finer point on the connection by postulating the equivalence between energy and information which has been recently experimentally shown to be the case. Today, information theory is used in compression, cryptography, climate science, and biological neural network inference.There are several people working in the Complexity Science Group at the University of Calgary on applying information theory to network inference of biological neural networks.
Open Public Data
I've spent far too many hours grinding through PDF transcriptions of financial reports, legislation, and election summaries to not appreciate government open data projects like the Canadian Library of Parliament's LEGIS database and Statistics Canada's Canadian International Merchandise Trade Database. These types of open, digitized databases are becoming more popular and necessary as government activities develop in size and complexity. We've already seen plenty of public frustration with government inefficiency. I enjoy implementing tools for analysis of these data sets. Some of my work may be posted here.
Financial Market Analysis
Market analysis is especially fascinating due to the high resolution and quality of the available time series and also because of the complexity of the system. In the past I've written artificial neural networks to discover correlations between stock price movements based on their histories of volume, closing price, spread, entropy, and variance but these efforts have failed to produce reliable techniques for forecasting. I'm currently interested in applications of information theory to understand the causal connections in securities markets. Unfortunately, intra-day data sets are expensive to acquire.
Telescope Design and Construction
Around the time I transfered my degree from geomatics engineering to astrophysics I became engrossed in the amateur telescope making community. The combination of technical theory, construction, and creativity is wonderful. Grinding glass is slow, methodical work but the payoff of seeing progress is well worth it. The work also creates a special connection with early scientists as the techniques used are still very much the same; there's nothing to it but to put your head down and push glass. My current project is a 20.3 cm (8") F/9 Newtonian primary. I also have a 40.6 cm (16") glass blank waiting to be rough ground.
Film and Television
Given that films have the influence they do over political discourse and societal moods I think not engaging with media on a critical level would be a mistake. People watch films and, consciously or not, have their thought patterns cultivated by them. Then they go to the voting booth. Films also provide insight into the cultures that produce them. The latest rash of Hollywood super hero films has been an impressive display of Western self-reflection. I love discussing films and hearing others' thoughts on them in both a narrative and creative technical sense. I'll occasionally post interesting things I find here.