The 10th CDA Data Analyst Certification Exam ended successfully at the end of June 2019.

We interviewed several excellent candidates in the CDA certification exam Level 1 before, and shared their thoughts in preparing for the exam.

Today we bring a few candidates who have achieved excellent results in Level 2 big data and modeling. Among them are graduate students who are studying in the school, new recruits in the workplace, and experienced data workers. Then they have What kind of experience to prepare for the exam?

Let's take a look at their style below!

LEVEL Ⅱ Big Data Analyst

Wang Xing

I hope to continue to do something interesting, meaningful and valuable in the big data industry in the future.

1. Current work

I am currently working in Jiangsu Xinwang Videoconferencing Software Technology Co., Ltd., mainly engaged in big data analysis and product development related to telecommunications industry.

2. Opportunity to apply for CDA certification exam

At that time, our company encouraged employees to sign up for the CDA exam. After reading the outline, I found that the content of the CDA level 2 big data analyst's exam was very practical. It was very close to the big data knowledge often used in my current work, and I signed up decisively.

3. How did I prepare for the zero-based data science

I signed up on May 31, and it took about 4 weeks to prepare for the exam. The study plan is usually about 1-2 hours after going home from work every day.

First of all, according to the requirements of the syllabus, all the contents of the exam are systematically passed. For some content that is commonly used, such as the knowledge of the Hadoop ecosystem, a simple review is enough.

For content that is not deep enough to be understood, such as data mining and machine learning, look for relevant information and make breakthroughs.

4. What are the knowledge difficulties in preparing for the exam

The basic principles of Spark, RDD, GraphX, and MLlib are all difficult, especially the MLlib part, and have some understanding of data mining and machine learning.

5. Recommended books and courses

The first is the "Hadoop Definitive Guide" that accompanies many people into the Hadoop world. This book is also a must-read item recommended by the CDA exam. The latest version is currently the fourth edition.

In addition, it can be combined with Professor Lin Ziyu's video course "Principles and Applications of Big Data" for better results.

If you do n’t feel addicted, I would like to know more about it. I recommend "Hadoop Application Architecture", which stands at the height of the architecture and explains in detail how many tools in the Hadoop ecosystem work together to form a complete construction plan for various big data analysis scenarios.

The Spark learning recommendation is also "Spark Rapid Big Data Analysis" by O'Reilly. The book gives a detailed description of Spark's architecture and related components, and is also a basic guide to enter the field of Spark analysis. It's just that the book was completed earlier, the Spark 2.x version has not been released, and part of the syllabus has not been covered.

Therefore, it is recommended to read the "Spark Programming Basics" edited by Professor Lin Ziyu at the same time, combined with the supporting Professor Lin's video course "Spark Big Data Processing Technology", Professor Lin's books and courses explain and describe Spark MLlib in great detail and are worthwhile one look.

6. Suggestions for candidates

For students who are engaged in the big data industry, especially Spark-related work, because the test content is quite practical, many of the test preparation content is the usual work content. So don't be too nervous, follow the outline, review the content you mastered, focus on what you don't understand enough, grasp the key points, and make breakthroughs.

For students who intend to engage in big data analysis, CDA is also a very good opportunity to enter the world of big data. It is recommended to systematically and in-depth study according to the books recommended in the syllabus. Know what it is, but also know why.

Also, be aware that big data is a highly practical industry, and you must do more while learning theory. From the basic Linux installation and Hadoop distributed construction, to the deployment and use of Hadoop ecosystem Hive, HBase, Flume, Spark and other projects, you have to try it yourself.

7. Future career development plan

Hope to work together with company colleagues to do something interesting, meaningful and valuable in the big data industry.

LEVEL Ⅱ Modeling Analyst

Guo Chang

The theoretical knowledge learned in school is often not enough. I want to strengthen my learning by passing the CDA certification exam with my skills.

1. Opportunity to apply for CDA certification exam

I am a graduate student in Applied Statistics at Anhui University. My undergraduate is also a major in statistics. From undergraduate to graduate students, the statistics major is becoming more and more popular. In today's statistics are becoming more and more popular, my ability must match the development speed of the major. The theoretical knowledge learned in school is often not enough, so I want to be close to my skills and strengthen my learning through textual research.

Since I was in the first half of the semester, I was exposed to data mining and the tutor is also the direction of data mining, which led me to be very interested in data mining and machine learning. I just learned that the focus of the CDA Level 2 modeling analyst's exam is on data mining algorithms and software implementation. Fuck, so decided to take the exam.

2. How did I prepare for the exam

Since I am a student, I have plenty of time. Considering that my proficiency in software is not enough and whether CDA L2 can pass depends on the actual operation. Therefore, the first semester of March is the first time I invest in software learning.

(I chose R with a certain foundation for R software) My preparation for the exam is divided into four stages:

1. Language (software) learning

The importance of practical operation has just been mentioned, choose a data analysis software or a language (R / Python) to ensure proficient operation. This stage needs to run through the entire preparation period.

2. Study of the key points of the syllabus

After I understand the syllabus, the focus of the examination is divided into two parts: data preprocessing and algorithm modeling. Since I am a statistical major, I deeply understand the importance of data preprocessing and have a certain foundation for this part of the syllabus knowledge point. I will focus on the algorithm modeling, look at the algorithm book to overcome the algorithm one by one, and the preparation time is about 1 month.

3. Check and fill in the gaps

After overcoming the difficult points, all knowledge points need to be checked for deficiencies. The best way is to read the book completely. I have read "Introduction to Data Mining" twice to supplement the knowledge points.

4. Sprint video

After reading the syllabus and textbooks, sprint the video to sort out the test points, pull the thinking back to the test line, make notes, repeat memories, repeatedly practice the mentioned operation questions, and prepare for the test. (Review time is 1 month)

3. What are the difficulties in knowledge preparation

For me, determining the threshold of the classification problem model evaluation is a difficult point. Because I have never been exposed to this kind of threshold adjustment problem in schools before, I learned the adjustment of the threshold for class imbalance problems through this exam, the threshold value is determined by the F value, and the threshold value method is determined by profit.

4. Recommended books and courses

Books should be close to the outline, I am mainly based on "Introduction to Data Mining", supplemented by "Machine Learning".

I learned about the CDA sprint before the exam. After reading the outline and textbook at least once, and then understanding the knowledge points, I will watch the video repeatedly and sort out the knowledge points. It will feel very clear.

Here are some of my notes:

5. Suggestions for candidates

1. Practice should not be taken lightly. The study of practice is a process of continuous accumulation and practice makes perfect.

Second, the textual research is not the purpose. It is the purpose to truly learn knowledge through textual research and to apply what is learned.

3. Since I also met many older sisters and brothers who were working and verifying during the test preparation process, there is not a lot of time for this type of candidates. It is recommended to choose a data analysis tool suitable for you.

4. Join the data analysis test preparation group to obtain the latest test trends and avoid independent combat.

6. Future development plans

Data analysis has become an indispensable skill. I hope to apply my knowledge of data analysis and data mining to my future occupations.

Yang Wenxin

As a modeling analyst, I want to help myself sort out what I have learned by passing Level 2.

1. Current work

I have graduated for more than a year, and I am a student majoring in statistics in the relevant direction. Now a modeling analyst, his daily work is related to data comparison.

2. Opportunity to apply for CDA certification exam

When I was in graduate school, I saw it on the forum of Economics and Management Home. I had the chance to report the Level 1 exam. This year happened to have a change in work. If I wanted to find a time to sort out the knowledge I had learned, I applied for the Level 2 exam.

3. How did I prepare for the exam

I started watching it in early June, about four weeks, the daily review time is about three hours after work and the weekend time.

According to the syllabus and analysis of the multiple-choice questions, I used "Data Mining" and "Introduction to Data Mining" to sort out the knowledge points; for the practical questions, I used R, and I practiced all the methods mentioned in the syllabus.

4. What are the knowledge difficulties in preparing for the exam

The part of the model evaluation involves a variety of evaluation indicators covering a wide range, many of which are relatively small, the relevant materials are difficult to find, the specific statistical caliber calculation method for each indicator is difficult to determine, and the specific test practice is also more flexible.

5. Recommended books and courses

Based on my own experience and learning process, I recommend the following books:

· "Introduction to Statistical Learning" is an ESL entry version, which can be practiced based on R language;

· Dr. Li Hang's "Statistical Learning Method", this year's second edition, the algorithm theory part is very thorough;

· "100-sided machine learning", covering the specific operations of most algorithms, including many application details.

6. Suggestions for candidates

The knowledge scope of the CDA level 1 and level 2 investigation should be no stranger to mathematics or statistics students. The exam itself is also a good opportunity to systematically sort out the professional knowledge learned.

The preparation of the exam is mainly based on the content of the syllabus analysis. According to the reference book, if you don't understand the content, you can turn to the Internet and read more books. Practical operation should pay attention to more understanding and practice in the early stage of data preprocessing and feature engineering, and not just limited to practical algorithms.

7. Future development plan

Do a good job at work, follow the leadership arrangements, pay attention to sum up the accumulated business knowledge; professionally improve the proficiency in tool use, pay attention to cutting-edge knowledge, and learn new algorithms and tools. Expect to be able to gain a foothold in the post and apply what is learned.

Su Jing

By preparing for the CDA certification exam, I hope to further establish my own data analysis and machine learning knowledge system.

1. Current work

After graduating from graduate school in 2006, he entered ZTE and successively engaged in router product software development, quality management and process improvement of various wireless 4G and 5G products.

As the company becomes more and more digitized, the full advancement of devops requires the mining of value from large amounts of data, product quality improvement, and process improvement.

2. Opportunity to apply for CDA certification exam

In my daily work, there are two main types of data to be analyzed. One type is project development process data, and the other is product KPI data.

From the earliest excel, to mintab, and 6sigma analysis methods, to using python to do a larger amount of data analysis, making big data analysis sustainable and automated embedded in the R & D process. In order to maximize the hidden value of data mining, it is necessary to systematically learn the data mining method of KDD, in order to more comprehensively abstract and build data models, on the one hand to improve the version stability of the product, to predict and fix problems in advance, on the other hand to make Manage to a new level and provide more reliable decision support.

Based on this goal, I hope to further develop my own knowledge system of data analysis and machine learning through CDA learning.

3. How did I prepare for the exam

1) The outline shall prevail, supplement knowledge blind spots, and establish a knowledge framework.

When preparing for the exam, you must closely contact the syllabus, check and fill gaps according to the content and knowledge points of the syllabus, and gradually establish your own knowledge framework. This is an indispensable prerequisite for preparing for the exam. Only after the foundation is laid well can we carry out further study.

2) Case-based, solve practical problems.

First learn Python language in combination with a small case, can realize the original chart made with excel through Python, with the Python foundation, combined with scikitlearn website to do complex algorithms and implementation. Because my goal is to solve the problems encountered in actual work, the learning process is based on cases, and certification is a natural process.

3) Take a multi-pronged approach, learn from the shortcomings and understand the differences.

Familiar with excel, mintab, spss, python and other ways to analyze the differences and pros and cons, choose the most suitable and fast way to improve efficiency, and deeply understand the pros and cons of tool differences

4. What should be paid attention to in the preparation

1) First understand the intent of the case question, first think about the business logic, then do data preprocessing, and select the appropriate model.

2) At the same time, it is necessary to be able to adjust the parameters of the model to ensure better processing results. After all, the case questions are scored according to the ranking. This requires in-depth understanding of algorithm principles and model evaluation indicators.

5. Recommended bibliography

· "Introduction to Data Mining"

· "Machine Learning"

· "Data Analysis Using Python"

· "Exam Preparation Manual"

6. Suggestions for candidates

1) The outline must be thoroughly understood, understand every word, every example.

2) It is easy to learn algorithms and models one by one, when to use which one, how to use it, and what restrictions should be understood clearly.

3) Data processing should be cautious and cannot be taken for granted. Remember the GIGO principle.

4) Sprint before the test, distinguish concept differences, and strengthen the simple hand calculation ability of the algorithm.

7. Future development plan

Through data mining and application of methods related to machine learning, more professional data mining and predictive analysis are provided, and product improvement programs and results evaluation are provided for different project demands. Let the data sound, generate value, and provide professional management and decision support.