Article content
Kristjan Vassil is Vice-rector for Research and Head of the Centre of IT Impact Studies at the University of Tartu. Thanks to advanced machine learning techniques and predictive modelling, he has developed a software which will contribute to the creation of the next generation of public e-services – from real-time economy monitoring to predictive medicine. What would be the impact of this tool on the industries and on the policy-making process?
Here is the full transcript of our interview.
The Estonian e-government ecosystem started to emerge since 2002, which is the date since when we have the digital ID available for the entire population. Since then, the number of public e-services has rapidly grown from almost nothing in 2003 to about 1,800 by 2016, they are provided by about 1,000 mostly public institutions, and they’re relying on almost 300 public data repositories. Public doesn’t mean that they are open to the public; public means that registers like vital statistics register, population register and so on contain information about the entire population for some specific statistical purposes. That as a whole has generated by 2016 a system in which there are about 600 million interactions occurring on an annual basis, which amounts to about 1.6 million queries per day in a population of 1.3 million people.
What is this all about when we talk about predictive public e-services? What could be the major potential applications?
The current e-Government ecosystem is based on the notion that we take a conventional public service and convert it into the digital realm, which basically is nothing else than taking a service and providing it online. You make things faster, cheaper, more accessible and so on, so all is good in that end. But one of the perhaps unintended outcomes of building the ecosystem of Estonian e-government was the fact that companies or public institutions started to accumulate data as an outcome of their normal day-to-day activities, that stand in different databases and they are not used for various analytical purposes. For example, the Estonian Tax and Customs Board collects detailed data about how each individual company in Estonia is performing in terms of their turnover and how much it pays taxes; from that information you can make inferences about how the company is doing, but those data are isolated in one data set and they are not used for any kind of analytical purposes. That [information] has opened up, and you see that virtually in every aspect of the Estonian e-government. Wherever you look that opens up opportunities on top of which you can start building what we call the second generation – or the next generation – of public e-services that use those data, and instead of simply converting the existing service into online realm, add value on top of those conventional services and we call those predictive e-services.
When we talk about adding value, what is this value that predictive e-services can have practically to some sectors of the economy?
Adding value, in those terms, means that we would not have been able to add that value in the conventional world without the existence of e-government. The specific example can be predictive medicine where we attempt to use the data that patients and our healthcare sector aggregate and we can start making sense of those data by issuing prediction scores about, for example, Type 2 diabetes. So, this is the example where I, as an individual and as a client of health care system, can be informed and take action based on the predictive kind of information that has been made about me by the government.
Would you like to explain how does the system work?
In Estonia, we have a little bit more than 160,600 corporate tax declarations provided by companies to the government on a monthly basis. In these declarations what we record is the information about their turnover, the income and labour tax expenses, and some auxiliary data. If you think about these data, when you aggregate them all up, you can start building very neat real-time economy surveying systems. For example, we have built an economic dashboard, which looks at how the economy is doing, how the economy is composed at the aggregate level in real-time. You can see the sectorial composition the economy and the fancy thing is that you can build several Key Performance Indicators (KPIs) on top of those characteristics and you can start monitoring them. With technology, you always have unintended consequences. One of the unintended consequences of this dashboard is the fact that it is not only useful for the policymakers, people from the Ministry of Finance people from the Central Bank or people from the Tax and Customs Board. But, if you think about it, in a way it ddemocratizes the access to economic information, and this information basically belongs to the civic education classes, to researchers and scientists who are deeply interested in how the economy is doing how it has been composed, of which components and so on and so forth. If you think about few interesting KPIs, I know, whether not too many Estonians know, that 80% of the total turnover of our Economy is being produced by less than 4,000 companies. There are more than 100,000 companies altogether, so these are the examples in which you can open up the information that the government has, and make it available for the public good so that you can increase or democratize the access to the information.
I think equity or equality in this term is really important: not only you can look in real-time how the turnover of the economy is performing, but you can also look at what is the composition of the labor market, which tells you the number of people who are employed at any given time; you see the impact of the economic global economic crisis of 2008; you see how the average salary has been increased. But, also, you see the gender and age composition of the labour market at the entire economy and at specific sectors in which you are interested in. This is all dynamically available to you. You can select those sectors that you’re interested in. Moreover, if we talk about equity we speak about the difference in the gender pay gap, this information can be followed and monitored in real time using data coming from Tax and Customs Board’s declarations, without compromising anybody’s interests directly. Estonia is known as one of the widest gender pay gap countries in the European Union, and we have set policy objectives that we want to achieve. But, we don’t have any tangible measurement how we would monitor our progress. By using these kinds of real-time economic dashboards you can go as deeply into specific KPI’s as you’re interested in, and we could achieve equality at a societal level with digital solutions.
