Next stop for Big data: fast data?

Summary: When all our behavioral data are networked, in the cloud, our next step should be what the "big data" can be analyzed to make predictions, but perhaps this is only a prediction, because "accidental", because of "people's thinking" and other reasons, many times, humans do not follow the common sense of cards, so, Can your data accurately predict your behavior? The answer may be yes, but not quite large data, but need fast data! Next stop for Big data: fast data? We encounter the following situations in our life or work: The company's goddess mm has always liked to eat Haagen-Dazs ice cream, almost every day to buy a cup, but one day, she took a DQ ice queen to taste with relish; company Cock Silk programmer Xiao Ming to work early, overtime, the completion of task code quality, corporate group building activities are also actively involved, Consecutive quarterly is the company's outstanding employees, suddenly one day, attitude resolutely proposed to leave, said to go home to teach. I have a friend in the big data service to predict and explain, friends say, if according to large data basic algorithm speculated that the goddess mm is not to eat DQ, because her behavior data has shown that she will continue to eat Haagen-Dazs; Similarly, the behavior of large data analysis, the program Ape Xiaoming will soon be promoted to a research and development manager or director, It is impossible to predict that one day he will return to his hometown to teach. So the question is, is it possible to build an application model based on large data of audience behavior to replace the real-time psychological feedback data? How are they combined? When the founder of the questionnaire came to the defenders, I threw the same question at him, which he said was a very interesting topic about the relationship between psychological feedback data and audience behavior data. He first popularized two concepts: one is big data, the other is fast data. For example, we are 11 shopping on the day cat or Jingdong, all of our actions on these sites: Browsing the web, comparing goods, placing orders, paying, evaluating goods, and so on, constitute a large data picture, and all the users on the cat's large data screen that constitutes a large data combination. The sky Cat can analyze which province's Goddess Cup is large according to the big data combination, predict which goods will sell well, or build a model based on a person's behavioral trajectory to predict what commodities she might be interested in and advertise for. When the user does not click on the ad or the user leaves the cat, one months after the cat again, we can not from the behavior of large data to find relevance or reason. This time, quick data appears, for not clicking ads or leave the cat one months before the user, we through a similar questionnaire feedback method (can also be through the Web site technology to get "feedback", reduce the threshold of real-time feedback), collect the user's ideas, based on this real-time data, We can know the user's psychological feedback in real time and take corresponding measures. To the defenders of the questionnaire network of users as an example, explained the audience psychological feedback interactive fast data application cases, such as the world's top 500 companies in the form of questionnaires to collect staff ideas, and combined with the day-to-day performance of employees to evaluate the assessment; the start-up company The Black Horse ticket collects the music by the questionnaire methodProducer's demand, rapid product iteration; Video tv with form for after-sales service advice collection and management; Millet company through fast data for intelligent hardware trial ... For why fast data can be used in many scenarios, to the defenders that, whether large or fast data, in fact, we in the application and interpretation, we can not leave the understanding of "human nature", the understanding of "human nature" is the key to establishing the model of data interpretation. And precisely at this point, real-time feedback interactive fast data can be more in the data reflected in the "human" understanding, because behavior is easy to lie, and attitude is difficult to deceive. When asked whether the relationship between large and fast data was replaced, he gave a negative answer. On the contrary, they can form complementary relationships, map and complement each other very well. For example, in the United States, when you visit the famous shopping website Amazon, on the one hand, it is based on your browsing behavior of large data referral books, on the one hand, when you leave the site will give you a 3-5-topic feedback form to understand your psychological activities, the next day you may receive its small gift mail or promotional mail. The creation of such large data and fast data interaction models has increased Amazon's satisfaction by up to 1.5%, according to former Amazon data scientists. Will fast data become a new hotspot after big data? Businesses have a need for large numbers of users, as well as more robust demand for fast-data interactions based on real-time feedback (even more valuable for consumer companies, fast data). In the United States, Surveymonky, a fast-data company based on questionnaires, has valued more than $2 billion trillion, and another fast-data firm, Qualtrics, has just completed a new round of financing with more than $1 billion trillion in valuations. Future, fast data, big data, one is more real-time feedback, one is deeper data precipitation, how will develop, let us wait and see.