News story 18th Feb 2016

Making data analysis easier

Most organisations collect data without having a clear idea of how 'big data' can help them achieve their strategic objectives.

Using data analysis to support decision making can give an organisation the edge over the competition, explains Professor Rob Hyndman from the Department of Econometrics and Business Statistics.

"Our primary goal is to make data analysis easier. We work with a range of organisations to get the best possible information out of their data. In addition, we provide a range of open source software applications for organisations to use," Professor Hyndman says.

Data analysis

Professor Hyndman's work primarily focuses on forecasting.For example, he recently worked with the Australian Energy Market Operator to assist them in producing long-term electricity forecasts.

"One of the questions we sought to answer was, how much electricity will we need on February 2035 around 4pm on a Thursday afternoon during an extreme heatwave event? We need to answer that question now, so that we can build the power generators to provide the electricity we need and to ensure we don't have black-outs."

The mathematical model that was created for this project is now used by all Australian states, as well as international governments, to assist in the planning of their future electricity generation capacity.

In addition, the team creates packages for R – a programming language favoured by many statisticians. These packages are programming tools that simplify the code necessary to complete common tasks, such as statistical modelling and data visualisation.

"Google uses my forecast package for R to predict the total volume of many millions of search queries. Yahoo uses my forecast package for tracking web traffic on millions of different web pages."

But it's not just the tech giants working with his software. Among others, The Guardian uses the package to forecast subscriber numbers by location and Nestlé uses it to predict the total demand for thousands of their products,

Professor Di Cook joined the Econometrics and Business Statistics department mid-2015, strengthening the Business Analytics team. Professor Cook's focus is on data visualisation. She is a sought-after speaker, and she has contributed to open source projects such as GGobi.

Pictures of data often provide the analyst with eureka moments, of discovering the unexpected. The unexpected can be errors in the data which if left unaddressed could invalidate the analysis, or genuinely unanticipated associations.

Di Cook

For example, in recent work studying standardised test scores of 15 year olds from around the world, she found that the gender gap in math is not universal. There are many countries where there is no significant difference in the average math scores, and there are five countries (of the 63 tested) where girls score significantly higher than boys on average. Australia is a country where there is a gender gap in favour of boys. On the other hand the gender gap in favour of girls is statistically significant in reading scores in every country tested.

Professor Cook was also the PhD supervisor of some of the field's "rock stars", including Hadley Wickham, "the man who revolutionised R", Chief Scientist at RStudio and an Adjunct Professor of Statistics at the University of Auckland. His work is used by organisations such as Google, Facebook and Twitter, as well as organisations monitoring the US election polls, just to name a few.

Hadley was also the keynote speaker at the 'making data analysis easier' workshop.

The event, WOMBAT (Workshop Organised by the Monash Business Analytics Team), was held at Melbourne Zoo on 18 and 19 February 2016. It has attracted some of the biggest name in statistics as speakers and was booked out within weeks.

For more information on WOMBAT, please visit the event page.