The big data analysis of animal husbandry, MUYUAN is doing this!

Introduction of MuYuan Foods Co., Ltd.

MUYUAN Foods Co., Ltd. is a national key leading enterprise in agricultural industrialization with the largest scale of intensive pig breeding in China, a large pig breeding enterprise with large-scale integration of self-rearing and self-rearing, and a large pig breeding enterprise in China. The company was founded in 1992. After 26 years of deve lopment, it has 88 wholly-owned subsidiaries and 2 shareholding companies. As of December 31, 2017, the company has the capacity to produce 10 million pigs per year, nearly 5 million tons of processed feed per year, and 1 million pigs per year. It has formed scientific research, feed processing, pig breeding, and breeding pigs. Expanded and commercial pig breeding as a complete closed pig industry chain. The company adopts a large-scale integrated breeding mode, fully realizes self-cultivation and self-support, establishes a food safety guarantee system and a traceability system, and realizes the whole process monitoring from site selection, raw material procurement, feed processing to pig breeding. It has ensured food safety and was awarded the first batch of “Henan Province Export Agricultural Product Quality and Safety Demonstration Zone” by the Henan Provincial Government.

Official visit address:http://www.muyuanfoods.com/

Background of the project

The data center of the enterprise technology department of MuYuan Foods is responsible for all the informatization work of the docking group. The business department involves the production department, the veterinary department, the quality control department, the finance department, the sales department, and the human resources department. The business department has relatively high requirements on the immediacy of the data. For example, it is necessary to adjust the feed indicators and timely adjust the feed to make the ratio of the feed more in line with the actual situation.

With the advancement of the times and the continuous expansion of the company’s scale, enterprise data and informationization work must be efficient and fast. Traditional data processing methods have been unable to meet the growing demand of data processing. The data analysis needs of the business department often change. The efficiency of the above-mentioned cooperation methods is not mentioned underground, and the needs of the business departments cannot be quickly responded. The current situation of high cost and little effect needs to be changed.

In order to reduce the development of the IT department and meet the flexible and varied needs of the business unit, the data center decided to purchase tools to alleviate or even get rid of this situation. Comparing several XXBI tools, it decided to purchase the soft** Finereport+FinBI **products as a business intelligence analysis display platform in the entire data analysis system. Complex fixed reports and highly flexible analytical reports are developed by IT departments, and BI self-service analysis is handed over to various business departments. Collaborate to handle all presentation analysis needs.

The following scenario cases are shared by the self-analysis system built by FineBI.

Overall architecture

• 1. Project plan

The performance limitation of the traditional relational database makes it difficult to support multi-dimensional query calculation of a large number of levels of data. In this case, if the traditional relational database is directly connected to the data analysis query, the performance bottleneck is prone to occur. Therefore, the project uses the data engine that FineBI needs to do data extraction. Due to the small amount of data in the previous period, according to the recommendation, the lightweight local mode is directly used, and the data can be extracted into the local disk and stored in a binary file. The multi-thread parallel computing is performed during the query calculation, and the available CPU resources are fully utilized. Therefore, in the case of a small amount of data, the display effect is excellent. Putting it together with the web application is very lightweight and convenient.

After extracting the data, the engine is similar to the data mart in the architecture, extracting the existing business data into the engine for storage, and providing query support with extremely fast multi-threaded computing. At the same time, for real-time data tables, real-time query can be directly connected. Real-time data and extracted data can be displayed on the same **DashBoard **page at the front end, and can be flexibly switched and debugged, which is convenient for users to analyze data flexibly.

The actual maximum single-table data volume of the bottom of the MuYuan Foods is hundreds of millions. For the analysis of the large amount of data (the data volume is about 5kw), the query of the database takes 10 minutes. After extraction, it can be displayed quickly within 3s, greatly improving the analysis efficiency of users.

• 2.Analysis mode

The science and technology department of MuYuan Co., Ltd. prepares data, preprocesses the data and assigns relevant permissions. Based on business self-service analysis, each business unit uses the department to view data, assist decision-making, and report to headquarters. The headquarters monitors the analysis of each department to grasp the business trends and assist the group decision-making.

For the underlying data management, the business package classification management mode based on the business theme provided by FineBI products is managed according to the department and the department. On the other hand, for enterprises, if they only use the information center to release pressure and decentralize the data to the business department, once the mouth is opened, the data gap will flood like another flood, and then it will go to another extreme, resulting in decentralized data. Management confusion, data caliber is not uniform, data barriers between departments and so on will arise, and this will bring great data security risks to enterprises. In addition to providing users with a data self-service analysis platform, FineBI products can strictly control the rights management. By classifying all users according to roles, only the template permissions are assigned to the users, and finally business analysis users can only see the data under the corresponding role, and the authority to manage the data, only assigned to the Ministry of Science and Technology, to achieve strict control of the distribution of data permissions.

• 3.Effect display

(1) Environmental Protection Analysis — Flexible and autonomous ad hoc analysis By screening the company’s venues in different regions according to the time, look at the recent sewage discharge and returning the number of fields, refine the analysis in abnormal time, see the details of the detailed field, and then summarize other information to locate the cause of the abnormality. Make further actions to reduce the amount of sewage discharged and increase the number of returning fields. The analysis in the figure below is a common kanban for environmental analysis summary data, and the data of these common indicators are viewed and judged according to the summary of the situation.

(2)Analysis by the Veterinary Department — Instant and accurate indicator monitoring This common analysis is used to query and analyze coughing conditions in different periods of different dimensions. Judging the health status of pigs in the near future from the overall trend, and analyzing the health status of the whole pigs in the near future and refining the problematic columns, so as to timely and accurately correct the symptoms and take corresponding measures to improve the health status.

(3) Financial analysis — knowing the financial situation of the company

In the common analysis, on the one hand, the data analysis of the materials out of the warehouse, through the time, material categories, company name screening conditions, quickly understand the material out of the company, and then combine the original inventory, so that timely Timely replenishment of various materials. On the other hand, it is mainly the food purchase situation of enterprises, and the procurement situation of each branch company is displayed. Understand the feeding costs of each branch, and make timely adjustments to the reasons for the over-priced branches to make corresponding adjustments.

Current Situation and summary

At present, the number of BI editors and analysts has reached 150, the total number of users used is about 3,500, and the average daily user usage is 1,500 times per day. The following figure shows that the number of users is small. At present, mobile applications such as Enterprise WeChat are also widely used internally.

Now basically every business department has gradually cultivated a group of intermediate docking personnel, who are all aware of business, and mainly do business-related matters on a daily basis. Some business units have a main job or a business, with BI analysis. Some departments’ dockers specialize in visual analysis and business data analysis.

Therefore, the IT department now only needs to prepare the basic data, pay attention to the use of the docker, and then communicate and guide the actual business with the docker. It greatly saves the labor and time costs of the IT department and brings real value to the business. For example, the veterinary department analyzes the body weight of piglets of different ages and finds the abnormal situation of individual pastures. Combined with the unified analysis of the corresponding pasture vaccine diseases, the problem is found and the economic loss of a pasture is avoided.