Shulou(Shulou.com)06/03 Report--

What should I do to build a user profile with big data? Many people do not know much about it. Today, in order to let you know more about big data's method of building user portraits, I have summarized the following contents. Let's look down together.

Entering the era of big data, one of the concepts we often talk about is user portraits. The use of user profiles in the Internet can achieve the commercial goal of accurate marketing, so this is why it is so important to build user profiles in this traffic-oriented era. If any company's products want to do a good job in fine operation, they need to build a user profile of the product and service. Let's take a look at the concept and construction method of user portraits.

I. the concept of user portrait

We often talk about user portraits. What exactly is user portraits? To put it simply, it is the tagging of user information. We use big data technology to collect users' social attributes, consumption habits, preference characteristics and other dimensions of data, and then describe the user or product feature attributes, and analyze and statistics these characteristics, mining potential value information, so as to abstract the full picture of the user's information. Therefore, the importance of user profile is self-evident, it can be regarded as the foundation of enterprise application of big data, it is the prerequisite for targeted advertising and personalized recommendation, and lays a foundation for data-driven operation.

Big data technology has reached a mature stage for a long time, and its application has become an indispensable and important part for Internet companies. From infrastructure construction to application level, there are mainly data platform construction and operation and maintenance management, data warehouse development, statistical analysis of upper applications, report generation and visualization, user profile modeling, personalized recommendation and precision marketing and other application directions. One of the important practical significance of user portraits is that it can really help big data's technology to really land in the field of marketing.

Second, the construction of user portraits

1. Process template

First of all, the construction of user portraits should be based on a full understanding and clear planning of the general direction of user portraits. only by defining the direction can we do a good job in project scheduling and personnel investment budget. This is important for evaluating key indicators and key outputs at each development stage. Then it sorts out the index system according to the business line, including user attributes, user behavior, user consumption, risk control and other dimensions. In addition, tag-related data can be stored in Hive, MySQL, HBase, Elasticsearch and other databases, and different storage methods are suitable for different application scenarios. We should also remember the key modules of user portrait engineering, including the development of statistical, rule, mining, and streaming computing tags, as well as the development of crowd computing functions to open up access between portrait data and various business systems. provide interface services and other development content. Then we need to iteratively reconstruct and tune the developed script. Finally, the application scenarios of the portrait include VIP service applications such as user feature analysis, SMS, email, internal messages, accurate push of Push messages, different customer service techniques for users, and extremely fast returns for high-value users.

2. Product function

For example, tag views and queries are mainly for business people. Users can hierarchically click on the tabs to see a detailed description of each tag. There is also a user crowd function for business people. When applying tags, it may not only look at the crowd corresponding to a tag, but also need to combine multiple tags to meet its business definition of the crowd.

3. Tagging

The construction of user portraits can actually be understood as "tagging" users. The way of tagging can be divided into statistical tags, rule tags and machine learning mining tags.

(1) Statistical label

Such tags are the most basic and common tag types. For example, for a user, the fields such as gender, age, city, constellation, duration of 7 days of active activity, number of days of active days of 7 days, and number of active days of 7 days can be calculated from user registration data, user access, consumption data and other fields. This kind of tag forms the basis of the user's portrait.

(2) Rule class label

This kind of label is generated based on user behavior and determined rules. For example, the caliber of "active consumer" users on the platform is defined as "≥ 2 transactions in the last 30 days". In the actual process of developing the portrait, because the operators are more familiar with the business, and the data personnel are more familiar with the structure, distribution and characteristics of the data, the rules of the rule class labels are determined through consultation between the operators and the data personnel.

(3) Machine learning mining class tags

This kind of tags are generated by machine learning mining and are used to predict and judge some attributes or behaviors of users. For example, judge whether a user is male or female according to a user's behavior habits, and judge a user's preference for a product according to his consumption habits. This kind of label needs to be generated by algorithm mining. In the practice of project engineering, the labels of general statistical class and rule class can meet the application requirements and account for a large proportion in the development. Machine learning mining tags are often used to predict scenarios, such as judging user gender, user purchase preference, user turnover intention and so on. Generally speaking, the development cycle of machine learning label is longer and the development cost is higher, so the proportion of machine learning label development is small.

The above is a brief introduction to how to use big data to build user portraits. Of course, the differences in the detailed use of the above have to be understood by everyone. If you want to know more, welcome to follow the industry information channel.

Welcome to subscribe "Shulou Technology Information " to get latest news, interesting things and hot topics in the IT industry, and controls the hottest and latest Internet news, technology news and IT industry trends.