The eighth CDA Data Analyst Certification Exam successfully ended at the end of June 2018. Recently, we interviewed several excellent students who ranked among the top in this exam. In the previous article, we interviewed the champions of Level 1 and Level 2 in the direction of big data, (the champion interview 丨 CDA exam is a powerful self-test), this In the article, I interviewed the top three in Level 2 modeling direction, so how do they prepare for the exam and learn?

Let's take a look at their style below!

In 2006 graduated from Shanghai University of Technology majoring in business administration. Currently working in HP's global spare parts supply chain department as a senior business analyst.

01 I am currently working for more than 12 years since I graduated from university. After graduation, he first joined Matsushita Electric China Co., Ltd. and worked in supply chain planning for more than 5 years. After joining HP's global spare parts supply chain department as a senior business analyst in 2011, he has more than 7 years of work experience. Work with big data often.

02 Opportunities for applying as a business analyst, based on my understanding of business knowledge, I have a certain foundation for data extraction, sorting, cleaning, analysis, and presentation of data reports, but the knowledge of mathematical statistics is relatively weak. The higher-level modeling-based prediction and classification problems in mining have not been systematically learned.

The training that passed the CDA certification exam not only re-learned the theoretical basis of data mining, but also re-learned the knowledge of mathematical statistics, and learned how to use Python to organize data, model, evaluate models, and let it serve as a business. As an analyst, I have also improved the skills of a data analyst, which will greatly help my future work practice.

03 How to prepare for the test efficiently For my review of how to pass the test efficiently, my summary includes two main points:
1. Deeply understand the theoretical knowledge of the syllabus analysis video. Don't memorize it. This is a systematic process of data analysis. It needs to be written down before and after, and combined with some statistical theory books to enhance understanding. Although the test is a level 2 modeling analyst, but also learn about the basic theoretical knowledge of statistics in level 1, can better understand.
2. Learn Python to lay a good foundation for getting started, and master the two basic packages of pandas and numpy. All high-level programming is based on laying a solid foundation of language skills. In the future, you can learn through some good websites, such as scikit-learn, to improve your programming skills.

04 Further understanding of data analysis

Through this CDA certification study, I also have a better understanding of business analysis and data analysis, summed up in three points:
First, data analysis is mainly to combine business needs, collate data analysis ideas and understanding of data. Language should be used as an auxiliary tool to choose a learning and easy to understand, such as Python. In addition, multiple languages have different levels of complexity in implementing the same function. You need to choose a language that is easy to implement. You do n’t need to stick to one language. I suggest that you can learn more languages.
Second, data preparation is the hardest, most time-consuming, and most impossible routine of all follow-up work (modeling, evaluation). Again, it requires an understanding of business and data to do a good job, which is the basis of all subsequent work.
Third, we must learn to look at the results (the results of model evaluation). The language already provides codes that are very easy to calculate various statistical indicators. , Is not blindly applied, so as to help you choose the most suitable model to get the best results.

Graduated from Tianjin University of Finance and Economics in 2015 with a major in statistics. After graduation, he has been engaged in data analysis. He has worked in consulting companies and the retail industry successively, and will continue this occupation afterwards.

01 Current work I am currently working as a business data analyst, in the retail industry, mainly responsible for the assessment of various indicators and analysis of company operations.

02 On the one hand, the opportunity to apply for the exam is to hope that you can improve your professional quality through the course and supervise your continued learning; on the other hand, you are also preparing for the transition, so you will first do some accumulation of knowledge, and the CDA data analyst course and certificate recognition And the gold content is relatively high. In the future, it will continue to develop in the direction of data mining and is expected to turn to the Internet industry.

03 How to prepare for exams efficiently
1. It is recommended to determine your career direction before applying for the course or exam. If there is no clear direction, at least make sure that you are interested in this matter, so that the later learning process and exam process will be relatively easy and pleasant.
2. Plan the time and learning progress in advance, thoroughly analyze the exam outline, master every detail and clarify the content of each model, and form a set of your own modeling system. Focusing on passing the exam and learning data mining, expand knowledge points based on time sufficiency.
3. Psychologically have the confidence to believe that they can pass, and have the patience to study and solve seriously in the attitude.

04 My test preparation plan, for reference: planning phase (2 days)
Preliminary preparations:
1. Interpretation of 4 outlines of books, it is expected to take 10 days.
2. Video assist, expected to be one month.
3. Check for missing vacancies, later practice, expected to be 20 days.

Summary: Plan your time and learning speed in advance, which will help you follow up later and save time. Preparation stage (55 days)

Stage 1 Carding (10 days)
Analysis points:
Carefully analyze the exam outline, clarify the exam content, and learn according to your own knowledge.
Contrast review:
The reference book is a complete study, complementing all the knowledge points with the exam outline.
Highlights:
Excerpts of incomprehensible, incomplete, and questionable knowledge points and mark important knowledge points.

Second stage study (35 days)
Video follow-up:
Compare the outline video analysis with the video to learn the specific content, pay attention to the details of expansion and key records.
The main points are perfect:
The main points of the first stage are to follow up, improve, complete and expand as the learning progresses.

Practice exercises:
Partitioning (objective question practice, description mining modeling, predictive mining modeling, model evaluation) exercises, each model at least three times.
Stage 3 consolidation (10 days)
Key memories:
Objective questions are memorized, real questions and simulated questions are practiced, and every detail is clear and understandable.
In conclusion:
Summarize the characteristics, advantages and disadvantages of each model, and use multiple methods to practice the same data to form a set of own modeling system.
Practical consolidation:
Practice the modeling operation repeatedly, improve the system, and try to adopt suitable and optimal modeling methods for specific data.
Exam process (half a day)
Take the exam calmly:

Answer the questions carefully in the order of the questions and complete the exam at your own pace.

Summary: master every detail, clarify the content of each model; take the exam as the center, learn data mining as the goal, expand the knowledge points according to the time sufficiency; be psychologically confident, believe you can pass, and have a serious attitude Patience to learn and solve.

Reference books:
Data mining: concept and technology. Machinery Industry Press. 2007
Introduction to Data Mining. People's Posts and Telecommunications Press. 2006
Statistical Learning Methods. Tsinghua University Press. 2012.3
Data mining in R language with business case studies. Electronics Industry Press 2017.9
Video assist:
R language data mining-L2 modeling analyst learning time is guaranteed 3-5 hours per day
Graduated from the University of International Business and Economics, majoring in finance. After graduation, he worked in an Internet finance company as a data analyst.
01 My current work is majoring in finance. I am fortunate to be exposed to many data-related projects. I learned a lot of related knowledge, including Excel, SQL, SAS and Python. I am currently working as a data analyst at an internet finance company.

02 Opportunity to apply
1. Data analysis This industry has developed rapidly in China and is also very popular. But generally speaking, there is no national unified examination, and CDA is recognized by more and more professionals and companies, so I chose CDA.
2. Everyone is lazy, and I am no exception. I want to pass the exam to more effectively urge myself to work hard, study well with a plan and evaluation criteria.

03 How to prepare for the test efficiently? I think that in the process of my study, there are three reasons why I have been supporting me:

1. Interests I am really interested in this position and industry. Whether it is to extract effective information from dirty data or to predict some complicated data, it will make me excited. Every small progress and achievement will make me feel particularly satisfied and grateful for myself in this industry.

2. Learn more, read more books, and practice more. Starting from preparing for the exam in early March, set a task for yourself every day, and review it repeatedly according to the outline.

3. Communicate the models and algorithms that have been learned with colleagues and friends, and help each other in the difficulties encountered in the project. Of course, they will also share their own learning experience and learn about some new developments in this industry. Now is the Internet age, effective information exchange will save a lot of time and will allow you to grow faster.

04 Future development plan Because my university major is finance, I hope I can better combine finance and data analysis. Provide more valuable predictions and suggestions for the company through analysis and modeling. At present, we already have the necessary knowledge and relevant project experience for machine learning, and we are constantly strengthening the skills of deep learning and artificial intelligence. I hope that I can provide more feasibility forecasts in the future, save more costs for the company, have a better evaluation of customers, and reduce customer churn.