Innovation in Statistics

Statisticians focus on solving clients’ problems and are constantly developing new tools for new problems. Sometimes this is because there are new questions to answer, new data not encountered before, and sometimes new technology has made new approaches possible. “Big data” is an example of this, where technology has created large new data sets and has simultaneously provided the computational power to analyse them.

Usually new methods are initially developed in specific application areas.  This leads to statistical methods that are associated with particular application areas, such as environmental statistics, quality management, epidemiology or genomics.  However, like all mathematical sciences, statistics distills ideas that can be applied anywhere.  No statistical method is really restricted to one area of application.  Some are simply not widely known.

Applying techniques outside of their original application area is at the core of Data Analysis Australia’s approach.  Consultants at Data Analysis Australia work across multiple areas and every project has several consultants, so the team will always bring broad experience from many areas.  Innovation means imaginatively applying what is already known.

Survival analysis was developed in the context of epidemiology, literally analysing how long people survive after a procedure or diagnosis. The problem is difficult since some people are still alive and it is not known how much longer they will live.  The analysis estimates how long they might survive given that they have survived for a certain time already.  Data Analysis Australia has recently applied survival analysis in energy modelling, understanding electricity consumption of large businesses by estimating what they might have consumed if they had not acted to reduce their demand.

Signal processing methods were developed for estimating time delays in radar and sonar arrays using multivariate spectra.  When faced with understanding a complex chemical plant where time delays occur between different stages, Data Analysis Australia used these ideas to develop an exploratory tool that could highlight time delays that suggested problems in the control system.

Transport surveys, such as the Perth and Regions Travel Survey (PARTS) conducted by Data Analysis Australia, make heavy use of itinerary information, giving not just where people go but also how they get there.  When asked by Tourism Research Australia to analyse the International Visitor Survey to understand what encourages visitors to travel around more of Australia, Data Analysis Australia recognised that this data also contained itineraries that had not been thoroughly treated as such.  Methods of modelling and displaying itineraries were transferred from the transport area to tourism, enabling novel insights.

Data Analysis Australia continues to develop novel solutions in a wide range of application areas. For more information see Our Services.

June 2014