Delegates from fifteen Singapore government agencies gathered for an OpenGov Breakfast Insight session on data analytics on October 26. Mr. Mohit Sagar, Editor-in-chief at OpenGov Asia, started the discussion talking about the need to put the power of data analytics in the hands of business users, not just keeping it with the IT department. It is about empowering the end-users.
Mr. Charlie Farah (above-right), Director- Healthcare and Public Sector, APAC at Qlik, talked about six universal trends in the context of data and the public sector. The first is rising costs, putting pressure on governments to find efficiencies in the way they are spending public money. Then there is a growing citizen thirst for open data. They want more transparency about how their government is working.
There is high impetus for cross-agency collaboration. Government agencies are breaking down silos and sharing information across government, finding links and connecting services together. Connectivity and digitisation are two other important trends, not just within countries, but with counterpart agencies overseas too. The final trend is governments using data for improving social well-being.
Mr. Farah said, “In an ideal world, it would be perfect to have all your data sitting in one beautiful data warehouse or data lakes and you can start doing your analytics and visualisations on top of that.”
But in the real world, data comes from multiple sources – operations, finance, workforce, supply chain. Then there is citizen data from multiple sources for governments. Either all the sources could be combined into a single data warehouse with enormous investments of resources and time. The alternative is to start connecting the existing data points from different sources and start gaining insights.
Today there is a lot of buzz visualisation. But visualisation is the last mile. It is about visual analytics, Mr. Farah said. Being able to connect data to see the story and making evidence-based decisions. The idea is to put the functionality and the tools in the hands of the people on the frontline.
Data at the Transport Accident Commission, Victoria
Mr. Bernie Kruger (above), Business Intelligence and Data Science Lead, from the Transport Accident Commission (TAC) spoke next. The TAC is a Victorian Government-owned organisation whose role is to promote road safety, improve the State’s trauma system and support those who have been injured on the roads, through an insurance scheme. The TAC is a ‘no-fault’ scheme. This means that medical benefits will be paid to an injured person regardless of who caused the accident.
TAC has a two-pronged 2020 strategy. It aims to reduce the number of fatalities on the road towards zero. And it aims to deal better with clients’ serious injuries, helping get their lives back on track.
Traditionally the focus was on the scheme’s liabilities. There were between 600 and 1000 rules to apply and check for with every single payment, to prevent fraud. Now the focus has shifted to the client. He said that it is all about the client and client-centricity.
Data insights will be a critical enabler for this strategy. TAC plans to establish an enterprise-wide approach to translating data. This will allow research and data to be shared across the organisation.
Mr. Kruger went on to outline the challenges faced in the use of data. Often there is inadequate buy-in from senior management. They do not consider to be an asset. It is rather seen as an operational tool. The value of the data is seldom measured. Organisations are not aware of the monetary value they can attach to certain data of a certain quality. Moreover, there is chronic underinvestment in IT.
In addition to poor data quality, there is often a lack of good data governance, further complicated with the move to the cloud. Who owns the data, is it business or is IT? Weak data lineage also leads to problems. For instance, Mr. Kruger said you tap into operational data and create a report. If the operation changes, how easy is it to alter your report?
TAC was using a specific software as a Swiss Army Knife, for everything from loading the data to cleaning, storage etc. But being married to one software like this can be a trap. The total cost of ownership (TCO) can turn out to be very high.
Another challenge is that the BI (business insights) team / data scientists are viewed as service providers in many organisations. This kind of culture hampers collaboration and results in an us vs them mentality.
How is TAC addressing these challenges? Firstly, a lot of data management activities are being automated, enabling far more focus on high value analytics and leading to better reporting and better insights. Everyone is taken along on the data journey, not just executives or IT. The benefits of data are shared. People are shown what is involved in a day for analyst and the challenges involved in extracting insights from data. Solutions have to be designed together with business. Data scientists should not just get the requirements and design the solution. Business has to be a part of the process. If everyone is not on-board, projects are going to fail.
Today everyone wants to jump on to advanced analytics, AI but the entire supply chain has to be dealt with. “We have implemented data science to quickly get benefits out of predictive analytics, machine learning, network analytics, geospatial analytics,” Mr. Kruger said. It is an experimental approach. Whenever anything works, its results are shared. Open source tools are used. For example, the data science team at TAC has adopted R and has been using it extensively. There are analytics pockets all through the organisation. To bring them together a Data, Reporting & Analytics competency centre has been set up.
Other considerations are the adoption of agile and design thinking principles, borrowed from the software development lifecycle, whether to have an enterprise data warehouse or a data lake (why not have both!), data discovery (allow the end-users to discover the data and its value) and cloud vs on-premise. There is still pushback against cloud computing due to security and privacy concerns. But the massive computing power available in the cloud is a major consideration for Mr. Kruger and his team.
TAC also extensively shares data with other agencies, which results in richer data and richer insights. For instance, if someone is in an accident, the ambulance picks them up. Their data is captured on an ipad. That data goes to the hospital. Before the person is even in the hospital bed, a claim is lodged on their behalf with TAC.
The Trauma Reception and Resuscitation (TR&R®) project is another one. It is a decision support system for the trauma clinicians regarding resuscitation of the patients and the relevant protocols. The system receives information from the ambulance, the vital signs monitor and displays it on Google Glass right in their field of vision. Algorithms prompt the Trauma Team in real time to confirm the state of the patient, perform procedures and administer drugs as well as assisting with diagnosing injuries. Ultimately the data is integrated back into the TAC, so that they can follow up on the claim.
Polling questions and discussion
When asked about top drivers for improving business information usage in public sector organisations, the response from the delegates was split between improving speed and accuracy of decisions, improving and optimising process performance, developing better policy/ products/ services and achieving better business transparency.
Mr. Chia Ti Yu, Director (Finance, Systems & Projects), Ministry of Finance said that the driver would vary according to the role of organisation and the individual. For his role, the primary use of data would be optimising process performance, for a more frontline role, using data might be about improved citizen interaction or developing better policy and services.
The inward and outward facing organisations have different priorities. For instance, an agency like GovTech (Government Technology Agency of Singapore) would be more focused on citizen satisfaction.
Around 56% of delegates rated their organisation’s use of data and data analytics tools as fair (“we use data in our decision-making process, but analysis is primarily a manual process), while 44% rated it as good.
The two biggest barriers identified by the respondents to integrating more data and analytics into day-to-day decision-making were the need to manually compile data from many sources and limited or no access to data. Significant amounts of data might not even be digitised.
Here also, the situation varies a lot between organisations. For example, financial data is completely digitised but some hospitals still use pen and paper for certain things.
There is also a culture issue. One agency tried to get business users to do more self-service. They experienced pushback. They didn’t consider analytics to be part of their role.
Mr. Sagar asked the delegates if their organisations’ management understand the value of data. Or is analytics considered to be an expense. For some, it is still viewed as an expense.
The agencies represented at the session were at different stages of their analytics journeys. Several are using a mix of manual tools and commercially available analytics and visualisation platforms.
In an area like health, at least 60-70% clean data would be required. A small difference in numbers can make a huge difference in health. If the data from two hospitals is not of similar standard, they can’t be consolidated or compared.
There could be different systems. There could be a lack of data definitions and standardisation. There could be issues regarding a ‘source of truth’, as in, when the same data is available from multiple sources, which should be considered to be the definitive source.
Then with user-generated data, every agency has a slightly different practice. Mr. Paul Loke, Chief Information Office at the Accountant-General’s Department – Ministry of Finance, said that you do not want officials to create a 100-line Purchase Order (PO) for buying laptops, but you do not want a single line item, showing ‘IT investments’. The latter provides zero visibility. It is about striking a balance between the two.
Mr. Loke added that while cleaning data, it is important to know the objective. If the end-result is a dashboard, then data has to be cleansed. But fraud or crime detection requires dirty data.
Mr. Farah pointed out that there will always be some data quality issues. Organisations should not wait to embark on analytics till they achieve 100% in terms of quality.
It is not just enough to get the right, cleaned data. The data has to be received in a timely fashion. Some organisations such as the Economic Development Board (EDB) have made the required investments and consolidated their data. Now users can go and pick up whatever they require. But with others, data is still in silos. Data stewards sometimes have a protectionist kind of attitude towards their data and behave suspiciously towards data requests. Frequently, once the request is placed, it takes a long time to get the information. Here, data governance is an area of concern. People are more comfortable sharing aggregated data.
And at other times, it could simply be a matter of not having enough time to respond to data requests, while running daily operations.
Challenges remain but progress is being made. All the agencies have at least made a start on their data journeys. Some have already laid a strong foundation. In others, pilot projects are demonstrating benefits and senior managements are gradually acquiring a better understanding of the potential of data.
Concluding the discussion, Mr. Farah said that data analytics will not provide all the answers on its own. It should complement human reasoning, enabling the people to ask questions of the data. It is about hitting that sweet spot between Spock’s pure logic and Captain Kirk’s human intuition.
