We live, reside and work in a data-driven age and our society is incriasingly influenced by the digital information technology. Smart sensors and mobile applications generate larger and more complex data sets, so called "big data". At the same time, large data sets are becoming more accessible in our daily lives through connected devices and ‘Internet of Things’ (IoT). Big data analytics can provide betetr personal services and address current societal challenges. For instance, to discover personal medicines against rare diseases by analyzing genome data. Another example is remote steering of drones to save people faster after an avalanche in hard-to-reach areas.

Despite these trends, realizing optimal value of big data remains a challenge. Mainly because the data sets are often complex, come from different sources and are difficult to interpret.

For example, a certain world population data set could contain various location data and air pollution data. The reason that these data sets are difficult to interpret is due to the fact that human intelligence, such as logical reasoning and domain knowledge, is often still missing in the data visualization process.

Granen mix complexiteit data

At the same time, there are privacy challenges with big data. One of the challenges is to design innovative big data applications, without compromising on privacy while maintaining individual control over personal data. Data must be made anonymized wherever possible. And when this is not possible, people want to stay in control of their own data and the added value needs to be clear.

Personal feedback and role of artificial intelligence

Often the user asks him/herself" "Why is my data collected?", but also "What is the added value for me?" End users need not only transparency and control, but also personal insights, specific feedback and suggestions. For instance, a simple message "The data has been sent successfully" or "your grandmother has not had a visit for 7 days, maybe you would like to visit her this afternoon?"

Both for privacy and quality assurance, and for the acceptance of big data and artificial intelligence applications, the end user must be more central. Interactive personal visualizations not only provide valuable insights into large amounts of data, but also create added value for users. Personal feedback makes the user more aware of how the smart algorithms work and how a predictive data model reasons about the resulting algorithms. Instead of asking complex consent questions and expecting the user to delve into how the system works, personal feedback offers transparency and support to people in making essential decisions.

Persoonlijk feedback

Different modern techniques are applied to translate big data into new insights. Multidisciplinary approach brings together user-centered design, informatics and various Artificial Intelligence (AI) techniques, such as machine learning and predictive modeling. Such approach enables translation of complex big data into real time interactive visualisations such as dashboards, story telling, serious gaming, personal coaching applications and decision making aids.

dashboard voorbeeld

For instance, smart analysis of historical data and real-time predictive modeling allows to predict when a calamity or epidemic will happen. Timely notifications can then help to prevent the following calamity. Artificial intelligence methods such as machine learning and deep learning make it possible to discover new new and life-saving insights that were previously not possible.

On the one hand, it is essential to be very aware of the risks of using personal data. On the other hand, by making the technologies smarter and more personal, the big data applications become indispensable to solve current societal challenges such as secure digital transformation by translating real-time data analysis from multiple sources into the optimal personal services, for the best price, best user experience and best quality.

Collaboration of various disciplines such as data science, computer science, interaction design, ethics, law, as well as of individual citizens with technology designers, public and private organizations is essential for the transparent application of smart self-learning algorithms. In addition to the importance of and the right to inspect how the automated decisions were made, the General Data Protection Regulation (GDPR) also sets requirements for the controllers responsible for the quality audit of the data sets, in particular requirements of lawfulness, fairness and transparency. Involvement of humans in the automated decision-making process remains essential to prevent the bias of algorithms (Article 22 of the GDPR).

Finally, in addition to valuable insights, personal feedback also offers other benefits for a sustainable and transparent application of big data analysis. In particular, by (a) creating more awareness about how the AI ​​algorithms work, (b) explaining how predictive data models reason about the decisions. AI ​​models can be iteratively improved by involving users the data analysis process and getting their feedback on the interactive representation of data.