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

This article mainly introduces "what are the commonly used big data analysis models and what are the characteristics of big data". In daily operation, I believe many people have doubts about the commonly used big data analysis models and the characteristics of big data. The editor consulted all kinds of materials and sorted out simple and easy-to-use operation methods. I hope it will be helpful for you to answer the questions of "what are the commonly used big data analysis models and what are the characteristics of big data?" Next, please follow the editor to study!

Nowadays, if we don't say a few words about big data, cloud computing, blockchain, edge computing and other terms, we will feel abandoned by the times. Well, today we will mainly explain to you what is big data and what are the common data analysis models when doing big data visualization.

So what is big data?

The concept of big data (Big Data) was put forward by John Masey, chief scientist of SGI, at the USENIX conference in 1998. He published a paper called Big Data and the Next Wave of Infrastress, which used big data to describe the phenomenon of data explosion. However, it was many years after big data really got the attention of the industry. Among them, big data's most important fermentation enzymes are GFS, MapReduce and BigTable published by Google from 2003 to 2006.

Big data refers to massive data or huge data, whose scale is so large that it can not be obtained, stored, managed, processed and refined in a reasonable time through the current mainstream computer system to help users make decisions.

What are the characteristics of big data?

Big data's 5V features are Variety (diversification), Volume (large quantization), Velocity (speediness), Value (low value density) and Veracity (authenticity). Among them, Variety means that there are many sources and formats, and the data can come from search engines, social networks, call records, sensors, etc., which are either stored in structured form or unstructured data; Volume indicates that the amount of data is relatively large, jumping from TB level to PB level. Especially in the era of mobile Internet, unstructured data such as video and language are growing rapidly; Velocity indicates that the data is timely and needs to be processed quickly and get the results, which is essentially different from the traditional data mining technology; Value represents a large amount of irrelevant information, which is of low value without processing, and belongs to low-value data. Veracity (authenticity) data quality varies greatly due to different data sources and recording methods, and this difference will greatly affect the accuracy of data analysis.

What is the processing flow of big data?

The general big data processing process has the following processes: data acquisition, data storage, data processing, data display. This is shown in the following figure.

In short, big data is a data integration with a large amount of data and a wide variety of data, which can not be calculated by conventional classification methods.

With so many big data, how can we use it?

After cleaning the data collected from different channels, it is time to apply big data. Define different data models according to our demand objectives, and filter the data through the data model to get the data we need. So what are the commonly used big data models in our daily work? Today we mainly analyze several commonly used models to make a brief introduction. For your reference.

1. Behavior event analysis

Behavioral event analysis method: as the name implies, it is mainly analyzed through the behavior of the event to obtain effective data. At present, it is mainly used to study the impact and degree of the occurrence of a behavior event on the organizational value of the enterprise. Then our enterprises can use it to track or record user behavior or business processes. For example, user registration, browsing the product details page, purchase, withdrawal and so on, through the study of all the factors associated with the event to explore the reasons behind the user behavior events, interaction and so on.

In daily work, operation, market, product and data analysts pay attention to different event indicators according to the actual working situation. For example, which channel has the highest number of user registration in the last three months? What is the trend of change? What is the amount of recharge per capita in each period? The number of independent users who bought from Beijing last week, according to their age distribution? What is the number of independent Session per day? Behavioral event analysis plays an important role in the process of viewing such indicators.

Behavioral event analysis method has been widely used because of its strong ability of screening, grouping and aggregation, clear logic and easy to use. Behavioral event analysis generally goes through event definition and selection, drill-down analysis, explanation and conclusion and so on. In particular, e-commerce companies through the early user behavior data collection, in the promotion activities can be purposeful regional, customized user advertising. Through accurate analysis of user behavior data, a higher conversion rate of accurate users can be obtained.

2. Funnel analysis model

Funnel analysis is a set of process analysis, which can scientifically reflect the user behavior state and the user conversion rate at each stage from the starting point to the end point. In fact, often used in business operations, the simplest should be our sales department's sales project funnel. The sales manager uses the project funnel to analyze the follow-up and winning probability of the next key projects. The sales funnel is also a data analysis model.

Funnel analysis model is also widely used in e-commerce platform. Mainly in the daily data management work, such as flow monitoring, product target transformation and so on. For example, in a product service platform, LVB users start from activating APP to spending, and the general shopping path for users is to activate APP, register an account, enter the studio, interact, and spend gifts. The funnel can show the conversion rate of each stage. Through the comparison of the relevant data of each link of the funnel, the problem can be found and explained intuitively, thus the optimization direction can be found. For the process analysis with relatively standardized business process, long cycle and many links, the problem can be found and explained intuitively.

3. Retention analysis model

Retention analysis is an analytical model used to analyze user participation / activity, which examines how many users who engage in initial behavior will follow up. To put it briefly, you held an activity and invited 1000 people to attend the meeting. in the course of attending the meeting, some people were not interested in the activity, so they dropped out of the event, and some users persisted. So must the persistent users be your target customers? Then it may not be right. We need a tool to identify the remaining users who are the real users. This is an important method to measure the value of products to users. Retention analysis can help answer the following questions:

Does a new customer become the behavior you expect to target for some time in the future? Such as purchasing behavior, participating in activities, etc.; a platform has improved the user's online experience and invited interested users to participate to see if there is a successful transformation?

4. Distribution analysis model

Distribution analysis is a classified display of the frequency and total amount of users under specific indicators. It can show the degree of dependence of a single user on products, analyze the number and frequency of different types of products purchased by customers in different regions and different periods of time, and help operators understand the current customer status and customer operation. Such as the distribution of users such as order amount (less than 100 yuan, 100 yuan-200 yuan, more than 200 yuan, etc.), number of purchases (less than 5 times, 5-10 times, more than 10 times) and so on.

Function and value of distribution analysis model: scientific distribution analysis model supports user condition screening and data statistics according to time, times and event indicators. Count the number of natural time periods (hours / days) in a day / week / month for people in different roles to perform an operation, the number of times to perform an operation, and the event indicator.

5. Click the analysis model

That is, a special highlighted color form is used to display the click density of different elements in the page or page group (pages with the same structure, such as commodity details page, official website blog, etc.). It includes the number of clicks on the element, the percentage, the list of users who clicked, the current and historical content of the button, and so on.

Click on the figure to show the effect of the click analysis method. Click analysis has the characteristics of efficient, flexible, easy to use and intuitive effect. Click analysis uses visual design ideas and architecture, concise and intuitive mode of operation, intuitive presentation of visitors' favorite areas, to help operators or managers to evaluate the scientific nature of web page design.

6. User behavior path analysis model

User path analysis, as the name implies, the path of the user's access behavior in the APP or website. In order to measure the effect of website optimization or marketing promotion, and to understand users' behavior preferences, it is often necessary to analyze the conversion data of access paths.

Take e-commerce as an example, buyers go through the process of browsing the home page, searching for goods, joining the shopping cart, submitting orders, paying orders and so on from logging on to the website / APP. The real purchase process of the user is an intertwined and iterative process, for example, after submitting the order, the user may return to the home page to continue to search for products, or may cancel the order, each path has a different motivation. After in-depth analysis with other analysis models, we can find the fast user motivation, so as to lead the user to the optimal path or the desired path.

7. User clustering analysis model

User clustering is user information tagging, which divides users with the same attributes into a group through their historical behavior paths, behavior characteristics, preferences and other attributes, and carries on the follow-up analysis. Through funnel analysis, we can see that the behavior of users is different at different stages, such as what are the concerns of new users? Under what circumstances will the purchased user pay again? Because the characteristics of the group are different, the behavior will be very different, so we can divide the users according to the historical data, and then observe the specific behavior of the group again. This is the principle of user clustering.

8. Attribute analysis model

As the name implies, users are classified and statistically analyzed according to their own attributes, such as checking the changing trend of the number of users in registration time and the distribution of users by province. User attributes will involve user information, such as name, age, family, marital status, gender, the highest level of education and other natural information; there are also product-related attributes, such as user resident in provinces and cities, user level, user first access channel source and so on.

What is the value of the attribute analysis model? The area of a house can not fully measure its value, and the location, style, school district and traffic environment of the house are more relevant attributes. Similarly, the user's dimensional attributes are indispensable to comprehensively measure the user's portrait.

The main value of attribute analysis is to enrich the user profile dimension and make the user behavior insight granularity more detailed. The scientific attribute analysis method can take "de-multiplicity" as the analysis index for all types of attributes, and "sum", "mean", "maximum value" and "minimum value" as the analysis index for the attributes of numerical type. multiple dimensions can be added, and the graph cannot be displayed when there is no dimension, and the dimension of the digital type can be customized to facilitate more refined analysis.

At this point, the study on "what are the commonly used big data analysis models and what are the characteristics of big data" is over. I hope to be able to solve your doubts. The collocation of theory and practice can better help you learn, go and try it! If you want to continue to learn more related knowledge, please continue to follow the website, the editor will continue to work hard to bring you more practical articles!

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.