I've been working in big data for more than five years, but if a layman asked me to explain to him what big data is, I'm not very good at talking about it. Can I say that is the massive data format, transmission, storage, query, display? is still too abstract. Can I say big data is a big data volume? Not necessarily, the data collected on a single machine may be several terabytes per day, but it's just a matter of monitoring the state of a machine. And the price of Apple a day in every city in the country may be just a few megabytes in size, but it is an example of big data.
The view of the book is very clear. The first is that the sample equals the overall. Before the big Data era, if you want to understand the situation of a certain market, it is generally used in the way of sampling survey, this way is inevitably biased, such as the people who cooperate with the survey, may be in itself biased. But in the big data era, we are directly oriented to the overall sample, can directly analyze the overall real situation, more objective. In the past there were two conditions that were not available, and one was the cost of data collection, which could now be obtained directly through the network. For example, people all over the country today are concerned about what, will be in search engine query records reflected. Second, the computing and storage capabilities are not available, thousands of high-performance servers can quickly calculate the results, in the previous calculator age is uncertain.
The second point of view is to correlate relationships without worrying about causality. The person who buys a thing is very likely to buy a B thing, maybe the two seem to have nothing to do with it, but even though we put them together, we are most concerned about sales, aren't we? Figuring out how a presentation might be easier, but trying to figure out the reason behind it, requires a lot of cost, and in this fast-changing era, it may be useful to use this correlation to generate value, leaving the rest to be analyzed slowly.
When I was reading this book, I was thinking about what the big data really was like. How is it different from the old times? I think that there is a wide range of areas, such as the beginning of the country, I said the price of apples in various cities, if you have such information, you can determine where the apple to make more money to pay, consider the longer-term is where to grow apples the most cost-effective. The book also lists an example of all flight fares and is similar.
In the big data age, I predict that the sensor field will be full of development, perhaps the sensor everywhere, we can obtain a variety of data through the sensor, based on these data to achieve some new value. Today's popular wearable devices are just a basic application of sensors. Google's driverless car is also an example of an application. But the sensor age I believe has not come, now is a brewing period.
Turning to Wenfeng--reading "The era of Big data"