The Difference between Data Middle Platform and Data Warehouse

In a sense, data middle platform is a kind of data warehouse, which is to extract data and establish a data warehouse. However, there are great differences in the data source, the goal of establishing data warehouse and the direction of data application between the two.
First of all, from the perspective of data source, the data source expectation of the data in the data middle platform is that the whole domain data includes business database, log data, buried point data, crawler data, external data, etc.
The source of data can be structured data or unstructured data. The data source of traditional warehouse is mainly business database, and the data format is mainly structured data.
Secondly, the goal of setting up the data middle platform is to integrate all the data of the whole enterprise, bridge the gap between the data, and eliminate the problem of inconsistent data standards and caliber. Data in the middle of the platform usually cleans the basic data from many aspects. According to the concept of subject domain, multiple subject domains based on things are established, such as user subject domain, commodity subject domain, channel subject domain, store subject domain and so on. The data middle platform follows three one concepts: one data, one ID and one service. That is to say, the data middle platform not only gathers all kinds of enterprise data, but also makes these data follow the same standards and caliber. The identification of things can be unified or interrelated, and provides a unified data service interface. Just like cooking, according to the standardized dish name, first prepare all possible materials. The traditional warehouse is mainly used for Bi reports, with a single purpose. Only the basic data is extracted and cleaned for the related analysis reports. When a new report is added, it needs to be done again from the bottom to the top.
Then, in terms of data application, the data application built on the data platform is not only for Bi reports, but also for marketing recommendation, user profile, AI decision analysis, risk assessment, etc. Moreover, these applications are light and easy to be developed quickly, because the important data analysis work has been completed and precipitated in the data middle platform, and the previous work results can be shared by multiple applications.
The traditional data warehouse is mainly oriented to reports, and the construction of data application is the traditional chimney construction, which is a development method from the beginning every time.
Finally, the data middle platform is built on the distributed computing platform and storage platform, which can expand the computing and storage capacity of the platform. Most of the traditional data warehouse tools are built on the basis of a single machine, once the data volume becomes large, it will be limited by the capacity of a single machine.
Data middle platform is not just a system or tool, but a functional department that provides data asset management and services for the whole organization through a series of platforms, tools, processes and specifications. Data middle platform is responsible for data collection, data asset processing and management, and provides data services to business departments and decision-making departments of the front platform. Therefore, the core of data middle platform should be data asset management and data enabling. Generally speaking, it's data magazine.