Part 1
For the last few months, I have been going through a quest called “to quit my position as a marketing analyst at FMCG and get an offer for a position as a product analyst in IT”. I would like to share my experience and systematize the information collected during this time from various sources. And in order not to be unfounded, talking about what skills a product analyst should have, I will start with a review of open information about the requirements for job seekers in hh.ru vacancies.
I parsed open data on vacancies posted on the website headhunter.ru on 10/28/2020 for the request “Analyst” and “Product Analyst”. The complete notebook and links to the data are posted here .
Before talking about my conclusions, I will make a small digression about the details of the analysis.
I took the list of requirements required for a specific vacancy from the Key Skills section of the position description. Not all HRs fill this field with high quality: some have made a seal (Note: "Phyton"), some have a
Perhaps, for a more accurate account of skills, especially soft skills (Note: “you have good communication skills to communicate with business and technical teams”), it is worth highlighting the requirements from the full job description and breaking them down into semantic groups.
Considering the above, I would consider the percentages in the tables below not literally as "the share of vacancies in the specialty where this skill is required", but as "the priority of this skill for a specific specialty."
In total, 1,178 ads were available for analysis, more than 60% of which are in 5 specialties: analyst, business analyst, product analyst, marketing analyst and web analyst.
How do the skills required for each specialty differ?
It can be seen that the key skills for each specialty are different: for a Product Analyst, technical skills are important (SQL, Python), for a Marketing Analyst they often mention marketing analysis and PowerPoint, and for Web Analytics GA and I. Metrica (
If we continue the list of top skills for a Product Analyst, it turns out that technical skills are followed by analytical (data analysis, analytical thinking, analytical research) and knowledge of statistics (mathematical statistics, statistical analysis, a / b tests, data mining). A complete list with the interpretation of the frequencies of the skills in the very first picture with the tag cloud.
What specialty is easiest to enter without relevant experience?
The easiest way is to look for a job at junior and intern positions in the field of data analysis in the specialties Marketing analyst and Web analyst - about 10% of vacancies are ready to hire people without experience.
More experienced people are most often expected for the position of Product Analyst: more than half of the vacancies are looking for a person with 3-6 years of relevant experience.
How does the salary differ by specialty?
There are few vacancies with open information on wages - only 63. Nevertheless, it is impossible to resist and not look at the distribution. For an adequate comparison, consider the median of entry-level salaries (expected work experience "from 1 to 3 years").
An average business analyst can count on 140t, a product analyst on 100t, and the least willing to pay is a marketer and web analyst: 60t. Marketers and web analysts, urgently learn BPMN or Python, SQL!
A couple of related articles before moving on to the second part:
- Analytics for analyst hunting - an overview of vacancies, skills and salaries from people who understand much more about HR things than I do.
- Text about the required skills at different grades in Yandex.
- An article on the types of analysts in IT (
read in the voice of Drozdov).
Part 2
The task of the second part: to collect in one place the resources on which you can learn data analysis for free, in particular product analysis. It will be useful mainly for beginners and those who do not want to spend money on a paid course.
By the way, if you are thinking about a paid course, here you can compare the formal characteristics (duration, price, level) for many schools. Unfortunately, the quality of the material and presentation is difficult to objectively assess, so you should look for reviews yourself.
Where to start learning Python:
- take a free introductory course from Yandex.Practicum on the basics of Python and data analysis.
Pros : the best start is hard to come up with: everything is simple, clear, interactive. And most importantly, by the end of the course you will have your first independent project and a rough understanding of Python's capabilities for data analysis ready.
Cons : there will be a desire to buy the rest of the course.
- look at courses on stepik, for example, this one .
Pros : You can brush up on your knowledge of data types, loops and see how to install Anaconda + Jupiter Notebook to start practicing on your own.
Cons : Very briefly covered numpy and matplotlib libraries and not covered at all by pandas.
How to start learning SQL:
- see this course on stepik or this one on kaggle.
Pros : Suitable for completely beginners.
Cons : these courses alone are not enough, much more practice is needed.
- solve problems in sql-ex , sql-academy or sql-zoo .
Each platform has its own pros and cons, it is worth finding the one that you like the most.
- it is for product analysis that familiarity with Clickhouse is often required.
Clickhouse has good documentation and you can practice it , but there is very little information about the nuances of practical use. For example, karpov.courses has a very useful webinar on calculating the retention rate in ClikHouse.
Where to start learning math:
- If it's really scary to approach, you can start from scratch at the Khan academy .
Pros : You can start learning math, even with addition and subtraction.
Cons : all videos are in English, plus the information is stretched, often you want to speed up.
Pros : everything is short and to the point and immediately with tasks.
Cons : you have to sit and decide, just like in school!
- Textbooks on theory. faith and mat. statue of V.E. Gmurman, S. Glantz.
Pros : everything is a little more stretched here than in the sources from the previous paragraph, but in more detail.
Cons : you have to sit and decide, just like at the institute!
- once again go to stepik for courses in theory. faith and mat. stat .
Pros : lecturers explain clearly, there is practice and discussion in the comments.
Cons : some points are greatly simplified, and some will have to be further figured out. But this is a minus of any MEP and self-education.
Where to get experience in Data Analysis and ML:
- — Python CSC, Open Machine Learning Course ODS .
: , , .
: ( ), . - , . - DA DS , , , — , , CSC .
- ML ML . .
:
- .
The first few lectures on algorithms and data structures in Python from MIPT are interesting to listen to, even if you are not going to become an algorithm guru and other gnomes: they are about how the Python
- Visualization (Tableau, Power BI, Y.Datalens, dash).
I Tableau has free public access, where you can learn for free to build dashboards. Participation in marathons from Tableau or Datalens can be very useful, if only for the sake of detailed training materials. If you want to complicate things: deal with dash .
- Practice-practice-practice.
Choose an interesting project for your portfolio on kaggle, there are datasets for every taste: from the classification of mushrooms to wine reviews and suicide statistics . For each dataset, there are notebook examples and discussions!
And if kaggle is not enough, collect data for analysis yourself from an area of interest to you - an example of simple site parsing in the first part of the text.
Basic recommendations:
- In order not to get confused and not lose motivation due to the huge number of tasks, it is worth making a list with priorities for each item. It was convenient for me to maintain a table on Google Drive.
- A good test of your own knowledge would be to read the theory and solve problems on some topic on one resource and immediately try to solve problems from a neighboring site on the same topic. In this way, gaps can be discovered and knowledge supplemented.
- Choose analytical publics / accounts to your liking in your favorite social networks: you will be aware of the news and there will always be something to scroll over a cup of tea.
- Keep in mind that when you finally find your dream job, this will not be the end of the journey, but the beginning! So pauses and quality rest are necessary in the process.