[In the data collection stage, the driver should evaluate the hitchhiker Qualification Access Based on more complete credit dimension data; in the data analysis stage, the analysis chain should be established between the key indicators and business impact]
Recently, two national data crises occurred successively, which showed that leading enterprises still have a long way to go to make good use of big data.
In the drip-by-drip homicide, the dilemma is that so much personal credit data fails to identify high-risk drivers. According to reports, the suspect Zhong had borrowed money from 51 institutions before the case, almost all the institutions that can borrow money, and there have been many overdue. If the individual credit data can be fully taken into account when examining the qualifications of a hitchhiker, those who frequently break their promise should be excluded.
In the case of the leak of the original data, the records of the opening of Huazhu hotels were leaked and put on the "dark net" for sale, and the privacy information of hundreds of millions of people was exposed. Huaju's dilemma is that it "doesn't know what's the use" of storing vast amounts of raw data from its users, and if it leaks, the company's image will be impacted and its share price will be at risk of plunging. Big data at this time has become "hot potato".
Hua Hua and Didi, according to the data scale standard, are "billion level big data company". But large scale is not equal to good use.
Such incidents occur frequently, only this year, there are video broadcasting site AcFun nearly 10 million user data leaks, the future worry about the suspected leak of 1.95 million user data. This is a wake-up call for the whole industry.
The author believes that to make good use of data, we need to make more systematic security settings in data collection, data docking, data analysis, data early warning and so on. For example, in the data collection stage, the driver should evaluate the quality of the rider based on more complete credit dimension data; in the data analysis stage, the analysis chain should be established between the key indicators and business impact.
These days, we have seen some changes, such as droplets in data early warning to make improvements, based on the position deviation information of abnormal data diagnosis; abnormal indicators, contact the background for alarm; once triggered active or automatic alarm, can be directly connected to the police for alarm processing.
Technical experts believe that if the above data processing in several stages, the security of multi-party block chain calculation, it will be able to make the problem better solved. For example, in the data collection link, block chain can enhance the authenticity of information and the enthusiasm of data sharing.
Zhang Jiachen, founder of PTS (Points), a large block chain data credit analysis company, told me: "A small number of key data can be stored on the chain, most of the original data can be stored under the chain. The authenticity of the information can be guaranteed by setting a checking mechanism in the block chain. That is to say, when there is a difference between the checking results of the information, the block chain can automatically judge and return the results. At the same time, block chain check mechanism encourages the correct information providers, and punishing people who provide false data. This ensures the enthusiasm and credibility of sharing.
In the United States, Uber is working actively with the University of California, Berkeley, to apply asymmetric privacy technology to user data sharing, thereby extracting user behavior data from the Uber system into useful insights into public governance and business as well as preventing sensitive information leaks such as user home addresses and travel habits. Personal risk.
Zhang Jiachen said the block-chain security multi-party computation is applicable to all shared economic platforms, such as drip, American Troupe and airbnb, because these platforms involve user privacy protection, pre-assessment of the credit and safety of service providers (drivers, riders, landlords, etc.) and real-time data in the service process. Analysis and early warning.
In the future, O2 companies like airbnb, pure online service Internet companies like Facebook, and traditional businesses like ICBC will benefit from secure multi-party data computing based on block chains.
Take the US Troupe as an example, in the stage of rider recruitment, we can make a prediction and analysis of the rider's life and borrowing status on his mobile phone. After analysis, we can score the risk of the rider and contribute the scoring system to the wind control system. In the case of airbnb, the same assessment can be done in the landlord assessment phase and the rental phase; such assessment does not require the collection of raw data, and all calculations can be done locally on the phone to maximize privacy protection.
Waonews is a news media from China, with hundreds of translations, rolling updates China News, hoping to get the likes of foreign netizens