Current location - Training Enrollment Network - Education and training - Simple user portrait analysis
Simple user portrait analysis
Portrait of users is to identify various characteristics of users, label users with various labels through identification, and then divide users into different groups through labels, and conduct product/operation operations for different groups respectively.

For example, Lamian Noodles said it was promoted on WeChat, because Lamian Noodles is a ready-to-eat food, which is more attractive to young people and more inclined to busy social animals in the city. Then Lamian Noodles said that the portraits of users are younger (age) and office workers (occupation).

User portraits have four kinds of labels:

Such as: name, gender, age, constellation, education, height, income, occupation, etc.

Such as: getting married, whether there are girls, whether there are boys, whether there are old people at home, etc.

Basic behaviors: registration time, source channel, last active time and last payment time.

Business behavior: Whether you have bought preferential goods or won excellent students, these signs are helpful for the later operation of the products.

This category is very different from other categories, just like the business behavior in the third category, which is the characteristic of business behavior, while business-related is to accumulate some data that other businesses will not record, such as sports and fitness products.

It will involve: fatness, height, body fat rate, body mass index, chest training or hip training, daily average 10000 steps, how many fitness programs have been collected, and so on.

1, registration information

At the beginning of registration, you are required to fill in basic information such as age and region.

And choose your favorite fields and interests, you search for keywords in the app and so on.

2, through the user's own existing functions.

For example: push from what you have bought, such as buying women's clothes+cosmetics. IP is often used to infer regions and so on.

3. Infer from the people around the user.

Transmission distance: Based on some attributes, there are people around you and users have a high probability.

Through behavior: through collaborative filtering, find the target users with similar behaviors.

Example: PivotChart of EXCEL

When you need to see which attribute, the chart will change according to your choice, as shown in the following animation:

At this time, you can clearly see the data of provinces and cities, or you can choose to pay more attention to these areas when advertising next time.

Because this function will be simpler, write it with a slicer next time.

There is also a dashboard that uses tableau, or python's pychart package, and I will teach you how to use it later.