Strictly speaking it's not what you experienced, but a project you studied. The company in question is not necessarily what we think is an Internet company, but perhaps after reading it, we feel it is. For a while because the work needs to learn big data processing analysis methods. Mainly the use of open source datasets, call python machine learning function package, follow some cases and textbook code to run the effect.

Brush a lot of data and code, most of the data sources away from their own lives is very far, such as a certain kind of rarely seen flowers, the price of housing in Boston, Germany's second-hand car prices, until one day brushed a famous real estate agency in the country's housing price dataset (referred to as a certain home, right).

This is an eye-opening experience, as a certain family has made a huge database of several thousand suites in several second-tier cities in China to train algorithmic models on price, location and other information. In addition to the well-known housing area, floor, age, several rooms, several halls, orientation, etc., other key data dimensions also include normalized housing and subway, hospitals, schools, large shopping malls and other functional building distance, the number of neighboring housing, the recent price fluctuations in the region, the characteristics of the suite to adjust the score and so on. Then with a not too troublesome machine learning algorithm training can be used to predict the price of various listings. The landlord has just given the listing information to the agent, even if the agent is a novice, a combing of input information can quickly get a very accurate prediction of the price information. Then with VR online viewing, visualization of the distribution of housing maps, etc., the big data and AI is not too raw and effective promotion to the front line of business.

This is the way to play, the other agents naturally can not play, a certain home has been equivalent to open a full map of the deployment, what place what conditions can be sold how much at a glance, low suction and high spit, the sand table to guide the roadside push; while others, especially those who are individual small agents, or a small form of checking the records and personal experience estimation, simply can not be compared to. This is also estimated to be the reason why a certain family can now easily merge a lot of seemingly small volume of intermediaries, all the way to expansion can not be stopped. Perhaps some of the less small agency owners are also attracted by these technology models do not feel the need to resist it. Of course, personal opinion may not be comprehensive, but the data set of companies that do a good job, this year is definitely not too weak.