In today's era of big data, the importance of data is becoming more and more apparent. However having data doesn't mean having everything. Many times big data can give conclusions but not explanations. Only by presenting the data through visualization and analyzing it can we find out the real answer. However there is more than one side to every problem, and when we face complex data problems, the core is to correlate the data.

After this, only then can we say that correlation analysis is performed. Correlation analysis (rules) that mining the phenomenon of association, from a large amount of data to find things, features or data between the frequent interdependence and correlation. Associative relationships include simple association, temporal association and causal association, etc., some of these associations arise from experience, but not always known in advance, but through the database of data association analysis, which has an important value for business decision-making, commonly used in physical stores or e-commerce cross-category recommendations, shopping cart co-marketing, shelf layout display, joint promotions, marketing, etc., to achieve the association of items with each other. Sales promotion and **** win, improve user experience, reduce the stocking staff and user input time, find high potential users.

How to quickly bring the data to do a consolidation, but also quickly presented, is very important. Data analysis and mining is very important, but sometimes some application scenarios and analysis and mining may not have much to do. For example, suppose a major earthquake occurs somewhere, in this catastrophic event, the most important thing is not to predict, but to diagnose, so that you understand what the situation is now, and in which direction the future is going to go. This requires the integration and presentation of data. The power of analytics is that you're able to analyze the entire process, not one part of it. You can see that the real story lies in quickly correlating different data sources.

In the past, we talked about big data analytics as encompassing data collection, cleansing, analysis and presentation, and today we talk about what seems to be correlation, analysis and presentation, so is it possible that the idea of big data analytics has changed? Indeed, it is a relatively large conceptual conversion. From the point of view of information, every data has value, the more data obtained, the better, some systems through cleaning or data processing, may remove some of the value. In addition, it used to be ETL, which is data extraction, transformation and uploading. Now the so-called transformation is done inside the correlation engine. The data is extracted, then uploaded, then transformed, which is ELT. ELT is much faster than ETL.