Become a better data analyst by avoiding these 3 things

Big data has come to dominate the world and analysing data is a skill sought after by many businesses, institutions and political parties as they try to navigate the future.

Collecting and analysing big data may seem straight forward but there are number of pitfalls. Not overcoming them can prove detrimental.

One of the biggest pitfalls is thinking everything can be solved with artificial intelligence. The disastrous handling of the Robodebt scheme in Australia is just one example.

Professor Dianne Cook is the program director for the Master of Business Analytics at Monash Business School. She says this is one of the most common issues with big data.

“When algorithms replace people there is a potential to run into problems. As a data analyst you really need to be able to detect any possible problems with the algorithms and assess ethical complications and understand issues related to fairness,” says Professor Cook.

"Society needs better data analysts, and for many more people to be comfortable working with data."

“That’s really what we mean by a data analyst. Having this type of diagnostic capability is crucial and is one of the things we teach in the Master of Business Analytics.”

The second pitfall is the time taken to get from original data to analysable data. A lot of universities teach students to handle data sets that are already in an analysable data form. But when you go to tackle a problem in practice, the data sets are often messy.

“When you start your first job it can be hard to know where to begin when you meet that gnarly raw data. If you haven’t got the skills and training, you could spend a lot of time and effort in this initial stage, getting the data into an analysable form where you can then apply your deep learning model or Bayesian network.

“We teach the critical skills needed to wrangle the data into the best form required for analysis,” says Professor Cook.

The third pitfall is a discontinuous workflow, where code and explanations get separated, often by switching software or doing manual cut and pastes. This has the potential to introduce errors to the analysis, or results that don’t match the explanation.

Professor Dianne Cook

The quality of your data analytics makes a significant difference to the decisions taken in different situations and the confidence you put into those decisions. Reducing the risk of errors is crucial. That’s where a reliable and consistent workflow comes into play.

“In the Master of Business Analytics we teach how to maintain a fluid workflow. We practice reproducible analysis and reporting, enabling tracing back through all the different steps taken in the analysis to detect when things have gone wrong, and also to make an easy update if new data becomes available,” says Professor Cook.

“This analytical ecosystem makes it possible to take messy raw data through to sophisticated modelling, and then to communicating the findings with amazingly glossy animated presentations, reports, web apps, and web sites. We use all open source and free tools, so once you leave and start working you will still be able to practice what you have learned.”

Why is it so important to not make these mistakes?

“Society needs better data analysts, and for many more people to be comfortable working with data,” she says.

“Our graduates are able to work in many varied organisations, such as Coles, Bunnings, the Salvation Army, SEEK, Deloitte, KPMG, Nous, or even in the Premier’s Office... the opportunities really are endless.”

Interested in gaining the skills to use data to better understand the world around you? Find out how the Master of Business Analytics at Monash Business School can open doors to new exciting job prospects.