Big data, big prediction?

On January 14, 2014, He Fan, deputy director of the Institute of World Economics and Politics of Chinese Academy of Social Sciences (CASS), acted as the keynote lecturer for the second installment of Hongru Forum in 2014, and shared his understanding on the way of thinking in big data era.

Why “big data” has become hot words?

The first reason is due to IT revolution. After IT revolution, the processing capacity, storage capacity and calculation capacity of computer data have been constantly improved. Compared with the speed of world economic growth, the growth of the stored information data is 4 times faster while the increase of computer’s data processing capacity is 9 times faster.

Secondly, data collections can be realized for more and more things, such as figures, texts, images, locations and social relations.

Three rules in the big data era

When data are big enough, we find that all social phenomena are statistical laws. In essence, we’re using statistic methods for data analytics.

The first rule: In big data era, we should explore the correlations instead of causality among things. Based on huge data, though we still need more efforts to understand their causal relationship, we are more likely to find out the correlations among things. As long as we can find the correlations, we may make the basically accurate judgments.

The second rule: consumers will suffer the thorough price discrimination. Under the statistic results of big data, merchants understand your behaviors better than yourself, so they can make accurate pricing aiming at each individual consumer and maximize the price discrimination. The more conveniences you get from the big data era, the less rights of privacy you can defend. Besides, in the future, it’s very likely that you can never get back these rights of privacy.

The third rule: information advantages of experts have been broken. Big data have completely overturned the situation of the information asymmetry among professionals and non-professionals, and therefore there are no experts in the big data age.

What can not be done by big data?

Big data may bring us blind optimism: we may believe that as long as there are data, anything can be predicted, but in the essence, the world around us is a complex system, and in theory, the complex system is unpredictable. The black swan events can not be predicted by big data. Besides, when we get more and more data, the proportion of noises in the information will become higher and higher, sometimes more data on the contrary may hinder us to make accurate judgments. In a word, big data can only let us be closer to the truth of the fact.

How to cope with the interlinked risks that are caused by big data

The benefits and risks brought by big data are interwoven together. In order to resist risks, firstly the independent operation of core businesses should be guaranteed; secondly, small risks should be tolerated and we should concentrate energies on preventing fatal risks; besides, we should have the safety island awareness and make well the isolation strategies for risks; finally, we should constantly maintain innovation.

In fact, when everybody is paying attention to big data varying at every moment, attentions should also be paid to slow variables. What are slow variables? Slow variables look like having no connections with you, but actually they have profound influences. They are the key for explaining the internal logic for events, but their shortcoming is that they have few data, and they are very far from the seeming direct connection of the true logic.