Risk Insights Blog

Looking for Something?

Search for posts and comments here.

file (1)

Analysing data doesn’t have to cost you a lot to generate deep insights

Updated: Jan 18



Analysing data doesn’t have to cost you a lot to generate deep insights

Two weeks ago, we presented at the National Investigations Symposium (NIS). We spoke about harnessing data to identify reputational and regulatory risk “hot spots”, and to help speed up and deepen investigations. A surprising takeaway for many attendees was that you don’t necessarily need to invest in larger data infrastructure at the outset. With the right frameworks, and a process that is scientific and defensible, you can harness your existing data to deliver insightful results that both illuminate and withstand query from senior executives, parliamentary oversight and the court of public opinion.


Since the Symposium, we have fielded many questions from attendees curious to learn more about our case studies and how to conduct similar data analysis. Some of the most frequent questions:


Q. If we can avoid investing in expensive data infrastructure and still get real benefits from our data, how do we get executive buy-in to become a data insights-led function?

A. Quick successes are the key. Focus on use specific cases – small, targeted data analysis projects with a clear impact. Take the example of procurement fraud:

  • Consider the tricky questions that data can help with (e.g. can we profile procurement activity for a certain individual / business unit? Can we geocode purchase activity?).

  • Identify relevant data sources that might prove useful (e.g. financial delegations registers, vendor master data file, accounts payable database).

  • Cleanse, blend/merge the data and analyse it – this is where domain / content experience is critical and machine learning can assist investigators reduce false positives.


Q. The case studies in the presentation all related to open data. Can you please provide an example of some of the open data you analysed?

A. For the procurement data case study, we extracted, combined and analysed four years’ financial data that can be accessed here.


Q. Are any regulators/integrity bodies publishing the results of their data analysis yet?

A. Some have started. For example, the Crime and Corruption Commission (Qld) has developed a useful corruption allegations dashboard that distils information about allegations received over the past three years. There is great opportunity for other integrity agencies to follow this approach.


The core challenges in delivering a successful investigation include reducing delays to finalisation, achieving completeness and coverage of evidence, and meeting stakeholder expectations. Data analysis can really help here, but it is only a tool (albeit a very useful one). As Chief Commissioner Hall from ICAC reminded us during the Symposium, ‘an investigator’s skills, experience and judgement will always be essential’. However, the right data, analysed using the right frameworks can significantly elevate an investigator’s capabilities, augment the exploratory potential of the investigation and accelerate its delivery. A compelling proposition to any senior executive.


#government #analytics #opendata

Public Sector Assurance / Audit Analytics