B2C website behavior analysis and data mining

At present, our B2C website has entered the final stage, and the next step will be transferred to the marketing and pattern improvement aspects. So when we are free, we will consider some other things out of the Code.

After the release of B2C websites and the completion of platform and resource work, the surprising success lies in the knowledge of the industry, consumers, and itself. Therefore, we need to slowly look at the essence through the phenomenon, in-depth analysis and mining of more things behind the website.

Here are some simple start points.

I. Requirement Analysis

Basic requirements:
Classified access information;
Product access and purchase information;
PV distribution;

Single point details. No association or cross-comparison is performed.
That is, a single element, such as the PV distribution chart of a specific time period, historical data comparison,
Classification PV ranking;
Same as above.
For example, find the popular categories and products in the second quarter to predict the sales promotion for the same period next year.


Intermediate requirements:
Association between categories;
Association between products;

Association. Find the trajectory rules of the content of the same nature.
For example, who cares about this category, what other categories do they care about? The association between the two categories is obtained.
Users who are interested in this content have developed into users who are interested in other content after several times;
In this case, we mainly analyze the relevance (track), that is, the time sequence of user access URLs.
Advanced requirements:
Association between users.

Personalized Association.
After data mining and analysis in the previous two stages, the relationship between users is initially analyzed. By analyzing the overlapping parts of user behavior, the same interests and habits of two users can be obtained to achieve personalized recommendations for users.