Dunstan M.

When I was young, I always fancied a career with lots of Mathematics application. No wonder, I decided to major in Mathematics in college and went on to join the National Statistical Office. The six years I have been a Statistician have been nothing but fun, I have enjoyed every part of the work, from the highly technical aspects (e.g., sample design and hypothesis testing) to the less technical ones (e.g., dissemination of survey results to the public and other stakeholders). I’m proud to have actively contributed to the successful implementation of large scale statistical projects (e.g., the 2008 Malawi Population and Housing Census and the 2010 Malawi Demographic and Health Survey). Due to my interest in the connection between Geography and Statistics, I was also given responsibility to manage the GIS section at the office. Four years ago, when I was considering undertaking PhD studies, I learned how Machine Learning weaves together Mathematics, Statistics and Computer Science. I instantly got hooked into this field and decided to explore it further by applying some of the techniques in my research. For my PhD research, I have access to a big data from a cellular company with information about user mobility. I have set up a Hadoop cluster to manage this data and also leveraged Apache Spark using both Python and Java to build predictive models in order to understand human mobility. Due to my Mathematics background, I have a knack to drill down to the theory behind any learning algorithm before I use it, as a result, I'm now at a point where I can suggest reasonable enhancements to some of the techniques. Now, I identify myself with the Data Science and Applied Machine Learning community. This is because I’m fascinated more with classification and prediction problems as compared to regular statistical analysis. As I'm nearing the end of my PhD, I’m really excited to take part in any initiatives to solve real world problems with Machine Learning.