Interviewer: Hello, we meet again, today we have invited Xu He to join our CDA holder interview.


Xu He: Hello, my name is Xu He, I was a control algorithm engineer before, and now I have transformed into an industrial big data analyst.

 

Question 1.

What do you think is the difference between a control algorithm engineer in traditional industry and an industrial big data analyst now? People think that the industrial big data industry is still relatively mysterious.



Xu He: The traditional control algorithm engineer mainly focuses on the control logic development and testing of the equipment management system. Then industrial big data analysts mainly use statistics, machine learning, signal processing and other technical means, combined with business knowledge to process, calculate, analyze and extract the valuable information and the laws of industrial operation process generated in the industrial process. The direct purpose of industrial big data analysis is to obtain various kinds of knowledge required for business activities to advance the refinement of equipment and enterprise management.


Question 2.

What are the main applications of industrial big data analytics in those areas?



Xu He.

Depending on the business objectives, industrial big data analytics can be divided into four types of applications.

1, descriptive analysis, used to answer the "what happened", the embodiment of the "what is" process. Generally, by calculating various statistical characteristics of data, various data are represented in a visual way that is easy to understand. The weekly and monthly reports and business intelligence analysis of industrial enterprises are typical descriptive analysis.

2. Diagnostic analysis: It is used to answer the question of "why did this happen". The key to diagnostic analysis is to eliminate non-essential random correlations and various artifacts.

3. Predictive analysis: to answer the question "What will happen? . For various problems in production and management, we predict the possible future results based on the factors that can be seen now.

4、 Prescriptive (guidance) analysis: It is used to answer the question of "what to do". To find the appropriate action plan for the problems that have occurred and will occur, and to effectively solve the existing problems to do a better job.

Different business objectives require different conditions, requirements and difficulties for data analysis. Broadly speaking, the four types of problems are of increasing difficulty. Descriptive analysis aims only to facilitate understanding; diagnostic analysis has clear objectives and right and wrong; predictive analysis, not only has clear objectives and right and wrong, but also distinguishes cause and effect and correlation; and premises analysis, often further combined with innovation in implementation tools and processes.


 

Question 3.

What analysis tools do you usually use most often when doing industrial big data analysis?



Xu He: The most commonly used modeling tool is Python, and the corresponding modeling process of machine learning and deep learning will be conducted for specific business problems.

 

Question 4.

Recently it is also the graduation season, what do you think about the employment prospect of industrial big data? What experience can you share with students who are looking for jobs?



Xu He.

The pace of industrial digitalization, networking and intelligence is accelerating, and the development prospect of industrial big data is also very promising. The national level proposes to transform from a large industrial country to a strong industrial country, and the government's policy formulation, driving orientation, development guidance and supporting support have also created an excellent environment for the development of industrial big data industry.

As for the suggestion, industrial big data is different from traditional data analysis, and also different from business big data analysis. The boundaries of industrial scenes are bounded by the mechanism of professional fields, and the analysis of industrial big data focuses on the integration of data models and mechanism models, and its important feature is the deep integration of data and mechanism. Therefore, it is recommended that students who are interested in industrial big data should not only understand big data algorithms, but also need to understand the general knowledge of the field, such as equipment operation mechanism, operation and maintenance process and other related knowledge.

 

Question 5.

We have talked a lot about your past and current career development today, what are your future career plans?



Xu He.

In the work practice to recognize the bottleneck difficulties of industrial big data analysis, often not the computer's ability to store and process data, but the complexity of the data association relationship. This complexity makes it difficult for some traditional data analysis methods to work, and cannot efficiently extract higher quality and more valuable knowledge. So to address this point, in the future, in addition to the ability to continue to strengthen the algorithm level, but also to strengthen the borrowing of engineering ideas and methods. Through the combination of technology-driven and business-driven, we will strive to become a technical expert in the field of industrial big data.

 

Conclusion.

Become an industrial big data analyst, industrial big data again on the wings of big data, accelerate the pace of digital intelligence of industry. Just now also talked about, want to enter the industrial big data to do data analysis-related work, not only to understand the data analysis-related knowledge, but also to understand the relevant knowledge of the industrial field. This is also an important tip for partners who want to do data analysis in the industrial field. Well, that's it for today's CDA licensee interview, thanks again to Xu He for the interview, thanks!