Eventually, typing questions into search engines will – to some extent – become a thing of the past. Not least because we are using our mobile phone more and more as our online communication tool.
It is increasingly common for people to talk to their phones, or to table-top loudspeakers with built-in microphones. This is because we want answers straight away, as we would in a conversation.
Statistics Netherlands (CBS) sees this development as significant; one that, in the future, may well become the most important way of providing society with factual information. We are currently developing a prototype computer-controlled dialogue application in which the answers are not limited to figures alone but can also feature data visualisations, infographics, videos, etc. By this means, we are offering our users a convenient way of accessing all of our output.
How does it work?
Someone might ask a question, such as: “How many men live in Utrecht?”
If the question was not entirely clear, then the digital assistant responds with a question of its own, such as: “Do you mean the province of Utrecht or the city of Utrecht?” The challenge is to be so flexible in constructing a dialogue that users are hardly aware they are talking to a computer. The system searches CBS’ public database for answers to people’s questions. This database contains every item of statistical data that CBS has ever produced and that is publicly accessible, also referred to as “open data”.
Our goal is to develop an application that answers users’ questions directly, by drawing on every item of content available at CBS. This also includes data that CBS collects for research purposes and for the purpose of creating open data, but which are not directly intended for publication. The need to rapidly formulate accurate answers while at the same time safeguarding the confidentiality of those who provide these data involves major technical challenges. Accordingly, another challenge has to be tackled first. This involves developing the technology needed to understand the questions and to link them to the existing open data. The first product version, based on open data, is being developed using technology from Google (Google Dialogflow and Google Assistant).
We look forward to hearing your views about this product and about its potential applications, and we are always open to ideas that might help us re-formulate the information dialogue. Please send us your feedback using the form below.