[CSDN editor's note] Many articles say that data science is the "sexiest" job in the 21st century, and data scientists can make a lot of money, so that data science seems to be a completely fantastic and wonderful job. But the facts indicate that data scientists usually spend 1-2 hours a week looking for new job opportunities. What is the reason for this?
Author | Adam Sroka Translator | Crescent Moon
Exhibit | CSDN (ID: CSDNnews)
The following is the translation:
Ideal career?
We often hear news about how popular data scientists are. News headlines such as "data scientist is the hottest profession in the 21st century" are endless, and the words data scientist are often seen on the annual high salary list.
Data science has contributed a lot. This is a challenging career. There is a lot of knowledge to learn, and you must stay energetic. Compared with other positions, data scientists have more autonomy to explore and solve interesting problems. In addition, in many cases, you will also have the opportunity to work with talented people in various fields.
Nonetheless, according to surveys, many data scientists spend several hours a week looking for new jobs. Among them, researchers engaged in machine learning spend the most time looking for new jobs. According to the 2020 Stack Overflow Developer Survey Report, this proportion is 20.5%, second only to academics!
If data science is an ideal career, then we have to ask: Why are so many data scientists looking for new jobs?
I used to face the same situation, so I hope to introduce my personal experience through this article and inspire everyone.
I worked as a data scientist for many years, and later moved to a startup company as a director. Now my job is mainly management and leadership. Since I have been a data scientist and managed a team (development team and data team), I can take a two-way standpoint and make some unusual points.
During my tenure as a data scientist, I also felt painful. Most of the time I worked in a startup company, and I skipped a few jobs early in my career. There are many reasons, but there are many factors that I have heard in other companies.
In this article, I intend to briefly summarize some common reasons and what factors I think can help alleviate this situation. My advice is mainly for the following readers:
● Data scientists who are not satisfied with their current job but cannot make up their minds to leave;
● Managers and organizations that cannot retain data science talents for a long time;
● Aspirants who plan to enter the field.
Before officially entering the discussion, I must first say that I still like data science. This is a very rewarding career, but you need to understand how to maximize your own interests.
Reality and expectations
The position of a data scientist sounds too powerful: the work is exposed to cutting-edge technology, solving problems in a variety of interesting ways, using new algorithms, and developing machine learning solutions that have a significant impact on the organization.
However, for many people, all this is just a dream.
According to my personal experience, as well as the experience of many people in the industry, reality and expectations are often very different. I have to say that this is the number one reason why data scientists feel frustrated and leave their jobs.
There are many reasons for this situation, but we should remember that it is two-fold.
Unrealistic expectations
In the early stages of career development, many data scientists have no experience working in a real organization. Just like the other people you see on social networks are not real life, people can only see the bright side of data science and think that this is the norm.
I found that people who have just graduated from university or left a research position in academia often have this wrong understanding. They will assume that there are no restrictions on time and budget. I often hear protests from data scientists, saying that they can't set a timeline for completing the work, they can only go with the flow. This kind of thinking is not correct, and it is not very consistent with the culture of most organizations.
You can only adjust the time according to the target range to be achieved, or adjust the range according to the time plan.
Another major reason for the departure of data scientists is that many of the jobs in this position are unremarkable. In most organizations, you not only need to do technical work, but also handle other boring tasks. If you don’t like to write reports, prepare speeches, toss over and over explain the basics of models or methods, manage projects, manage expenses, or try to win the support of other departments, then you will definitely run into trouble.
Ruthless reality
There is also a very common phenomenon that the infrastructure and data processing you expect does not actually exist.
Before, I worked in a startup company and I was the company's second data scientist. A colleague of mine has been working in the company for 18 months, during which all his time was spent building some basic data pipelines. Fortunately, they managed to persuade the relevant personnel to approve the budget after all the difficulties, and also solved the security and IT problems encountered when adopting the new cloud technology.
Sometimes, you have to be a smart technician who can complete the job smoothly even in the face of ambiguous requirements. And being able to do a good job in data science is second.
If the team lacks experienced data scientists, or the managers of the organization lack experience in managing data scientists, these problems will become more serious. If you are fighting alone, your views will hardly resonate. In this case, it is easy to lead to an unpleasant working environment and unfulfilled expectations.
As a data scientist, you might think that your job is to build intelligent models and get as much value as possible from the data. In the first few months of joining the company, you will be busy setting up the infrastructure and pipelines necessary to obtain data.
However, the company's senior stakeholders only see you spend a lot of time, but with little effect. In fact, if you can show some simple charts on the regular board meeting, they will be very satisfied. They only saw expensive resources being spent, but they didn't see you delivering value quickly.
This disconnect in communication caused disappointment on both sides.
If you have the opportunity, you should ask some relevant questions during the interview:
● Who at the top of the company decides to introduce data science?
● Do they have experience in data science? Or do they just follow suit?
● How many people are in the data team?
● Do you have a data engineer/analyst/operation and maintenance engineer? Or do I need to take on all the work?
The above content seems to be relatively negative, but the actual situation is not so bad. Data science work in many organizations is going well, but you need to balance your expectations with the actual situation of the position.
Wonderful office politics
When it comes to office politics, I believe many people will feel a headache.
An excellent team, excellent management, but depressed because of office politics, such stories are endless. Some people around me have this personal experience: the only senior leader of the data science department in an organization was forced to resign, and then the entire team was reorganized to take charge of some boring daily tasks. Everyone had a talent but nowhere to display it.
Office politics is an inevitable part of career development. You don't have to participate in this game, you can hop with your superb and rare technical skills.
I suggest you to learn about the "EVLN model" in organizational behavior (Exit, Voice, Loyalty and Neglect, that is, exit, voice, loyalty and sabotage). This model was proposed by economist Albert Hirschman in the 1970s.
When something goes wrong, you have four options:
● Exit refers to resignation, which puts the company in a worse dilemma, because the problem still exists, but you have lost your technical skills and experience.
● Voice refers to stand up bravely, discuss issues related to organizational change with the supervisor, and express one's own views, etc.
● Loyalty (Loyalty) refers to waiting patiently for the improvement of organizational change. If your patience is exhausted, but things haven't turned around, then you may choose other methods.
● Neglect refers to all kinds of passive sabotage, such as being late, leaving early, and using working hours to do personal things. This practice can easily lead to one's own expulsion.
Among the above four options, only "voicing" showed a positive side. However, if you choose this method, you are bound to fight office politics.
In many cases, the influence of office politics may even exceed salary packages. Your budget has been slashed, and you will definitely feel very powerless. At this time, you can weigh which way you should take. Sincere communication with a certain leader may become a catalyst for change.
If your organization is smaller and you are more accessible to decision makers, then I strongly recommend that you communicate with them. Unlike many people’s impressions, people usually want to do the right thing, and few companies will hire people who maliciously oppose you.
Generally, senior stakeholders may not have the opportunity to understand the needs of the data science team. You need to spend some thought to show them how to increase value. A good relationship with them can help them use your technical skills to get the most value, and it can also help you better understand what top leaders really care about.
In the early days of data science, I jokingly suggested to the CFO or financial director to automate some workflows. In this way, you can prove your worth to those who have the right to make budget decisions and win alliances for you. Of course, this is only a half-joking, because these people are very busy and are often overwhelmed by Excel.
You need to win the favor of people who are influential in business decisions. Most of them don't care about your algorithm at all, and they don't have the slightest understanding of statistics. You only need to help them complete some simple tasks, retrieve basic data, realize some automation or report work, and you can win their favor. Always smile and establish a good image. Over time, you will find that you can win the right time and place.
Data Science == Data Everything
Only by solving the problems of office politics can a good image be established. But this is a double-edged sword.
Many people don't understand (and don't care) what a data scientist means. As mentioned earlier, they only treat you as a smart technician who can handle everything. You can access all the data and are equipped with various technical tools. When they encounter problems, the first thing they think of is you.
If you can handle these issues well, it is naturally good, but it will become a burden. When people start to rely on you and put pressure on you, you will feel uncomfortable. Soon you will find that you are doing some junior DBA work 80% of your time.
I often tell companies that data scientists can do anything, but they are generally slower and more expensive than others. Our pressure comes from anything.
The skills of this position are very extensive, and the responsibilities are not clear, so it is easy to be forced to take on other tasks because everyone does not understand. You should communicate with senior stakeholders to help them hire DBA or BI developers, and you can free up your hands to do what you really want to do.
This method can also avoid being isolated. If your team of data scientists is very small and isolated, the professional skills you have about data can prevent yourself from being isolated. Data is your area of expertise, and others are not interested in competing for ownership of this area. You can help build better organizational structures, expand the responsibilities of data scientists, and expand collaboration with other teams.
to sum up
As a data scientist, it is not enough to master the latest tools and algorithms. Correct your mentality, don't have unrealistic expectations, and at the same time, through good communication, let your superiors understand this position, you are more likely to succeed.
I hope that the text will be inspiring for data scientists, organizations that hire data scientists, or people who want to get involved in data science.
Original link: https://medium.com/swlh/why-so-many-data-scientists-quit-good-jobs-at-great-companies-429ea61fb566
Disclaimer: This article is a CSDN translation, please indicate the source for reprinting.
Make an appointment for the live broadcast of "Big Coffee Is Coming", win gifts such as commemorative sweaters and mugs, and leave a message in the live room to ask questions. If the question is accepted, a mug will be given directly ! First come first served!
☞360搜索、UC浏览器等被3·15点名应用已下架；马斯克宣布通过NFT卖歌；美团App再发力社交，内测 “群聊”功能 |极客头条☞暗藏 15 年，Linux 惊曝 3 大 Bug 直取 root 权限！☞迷惑行为赏析：把 jpg 图片加密后卖了 247 万
☞微软高层人士变动！张祺晋升为微软公司全球资深副总裁