Any idea born in your mind is the outcome of data analysis. You have a problem => you think about it (analyze) => you come up with an idea of how to solve it. If you’ve analyzed the data properly, the solution will work. So, the essence of data analytics is to boost data-driven decision making. This is what data analysts do, in brief. If you want to know more, the best way is to ask an experienced data analyst. And that’s what we’ve done. Elvira, a data analyst at Railsware, agreed to answer a few FAQs about data analysis and explain her role in the Coupler.io project. Here we go!
What does a data analyst do?
Data analysts (DAs) research and interpret data to make it understandable for decision makers. They validate hypotheses or carry out A/B testing to find answers to emerging questions.
For example, there is a need to understand why the churn rate is growing. There is a hypothesis that users face an error, and hence churn. A data analyst needs to test this hypothesis (prove or disprove). Based on the outcome, the product manager will be able to make a better decision on how to increase retention.
In some companies, data analysts can be allocated to machine learning (ML) tasks, if the scope is not big. Usually, a good DA has the background to cope with ordinary ML stuff.
Which stakeholders do data analysts collaborate with?
The key stakeholder is the product owner. This is the person who has the principal vision of product development. Since I’m engaged in a few Railsware projects, I collaborate with different product owners.
But, mostly, data analysts collaborate with marketing teams. Marketing specialists request to analyze either a hypothesis of a specific activity, or the outcome of the approach they have used.
For some projects, data analysts participate in the content creation process. This mostly refers to knowledge sharing via blog posts, video tutorials, conferences, meetups, and other channels. For example, check out my recent blog post on the Railsware blog: How to Import Users’ Data Into Email Sending Platforms via API.
Why do companies need to analyze data?
Data analytics keeps your eyes open. This means executives understand what’s going on with the business/company/product and can react to issues in a timely manner. However, it is not only executives who will benefit from analyzing data. Users of a product or service are the ultimate beneficiary of all improvements you make after a proper data analysis. How is that?
For example, let’s say your churn rate goes up. A data analyst will analyze what makes a user leave and you’ll try to fix that issue. As a result, you’ll retain other users by making them happy. This has led us to the following conclusion:
Data analytics is not only about analyzing tables or databases – it’s the interpretation of information (numbers, comments, feedback, etc.) received from different sources. Customer development interviews, questionnaires, and customer requests can also be data sources and hence be analyzed.
What is the difference between data analysts, business intelligence analysts, and data scientists?
The main task of business intelligence analysts (BIs) is to automate reporting. They collect data and transform it into more easily usable formats, such as dashboards and reports that are used frequently.
Data visualization is an important part of a BIs’ scope. Business analysts build dashboards displaying versatile business metrics such as MRR, sales rate, total revenue and others.
Data analysts can use the data collected by business analysts. As a rule, large companies with many departments or products can own an army of BIs. If no BIs are in the company, data analysts perform their tasks instead: collecting data, automating data collection, creating reporting systems, building dashboards, etc.
Data scientists perform investigative data analysis. They not only verify hypotheses to find answers, but also build and deploy machine learning models to meet specific business needs.
For example, say you manage an online marketplace and need to remove any sales of illegal products and services (drugs, smuggled goods, etc.). A data scientist can use ML methods to build a model that will reveal unwanted sellers so you can ban them from the market.
The aforementioned rules are not as strict for every role. In some companies or on some projects, one person may consolidate all three jobs into one.
Do data analysts code?
Data analysts usually have programming skills in Python, R, and some other languages. Structured Query Language (SQL) is an essential, basic skill for every data analyst. We even released a SQL Tutorial for Beginners on our Railsware YouTube channel. For almost any project, you’ll need to make SQL requests to pull data from databases. So, yes, we do code. Open source and flexibility are a few benefits among many of using programming languages in our field.
However, using code for data analysts can be unnecessary or even forbidden. This mostly depends on the industry and country. For example, banks in Kazakhstan are subject to legal limitations on data processing and storage. As a result, only user-friendly analytics software like Excel or SAP are acceptable. The same is observed in the companies, which have purchased and incorporated specific analytical software.
Which programming languages/tools do data analysts use?
- Programming and database languages: Python, R, SQL, etc.
- Analytical tools: Excel, SPSS, SAP, STATA, etc.
- Visualization tools: Tableau, PowerBI, Qlik Sense, etc.
What are the soft skills required to become a data analyst?
Have you stumbled on a spurious correlation between Nicolas Cage and drowned people? The idea is that the number of people who drown in swimming pools correlates with the number of the films Nicolas Cage have been featured in. But that’s nonsense, isn’t it?
That’s where critical thinking comes into play. A data analyst needs not only to process data but also interpret it in the right way, which is more valuable than the data by itself. There may be a mathematical correlation between Cage’s films and drownings, but there is no logical correlation. Data analysts may find such correlation traps anywhere, so critical thinking helps to avoid them.
I call it a Google skill: the ability to search for data and information. Frequently, you get a very specific task that is poorly covered by any data analytics books or online courses. For example, recently I needed to parse Coupler.io positions on the G Suite Marketplace by specific keywords. I doubt that you would find any guide on how to do this, but you can create your own. To do this, you need to know how to frame a question to Google (or any other search engine) to get your desired results. It’s a very important skill to have.
Decomposition of a task means breaking down the task into smaller ones. For example, when I joined Railsware, my first task was to analyze one of our products – Jira Smart Checklist. It was a big task that included the analysis of sales, installations, conversion rate, and many others. A few questions arose, like “What to begin with?”, “Which priority to set up?” and others. To answer them, I built an implementation plan, where I decomposed the task into smaller pieces and prioritized them. This helped me to structure my scope and deliver the results in an efficient way.
Intelligibility of analysis
The outcome of data analysis must be clear and easy to understand for all stakeholders. Let’s say you’ve analyzed a certain issue and came up with a certain insight. It looks clear to you as a data analyst, but will it be clear to the person that will use it? You should unfold all the conclusions you’ve found during your analysis. If a product owner/manager, marketing expert, or another stakeholder was not considered into your concept, the job is not yet accomplished.
Desire to learn and develop
Data analysts do not sit on the sidelines. They constantly learn new tools, techniques and approaches. It’s quite important to grow as a professional. So, data analysts must be teachable; i.e., open and eager for new knowledge.
Education background: University degree or online self-education
You can find plenty of academic programs on data analytics in universities around the globe. You can go into an academic career as well. The healthcare industry is currently the most in-demand field for analytics. One of the best examples is IBM Watson Health. Data analytics, combined with machine learning models, help accelerate discoveries and amplify human knowledge.
Online options for self-education of data analytics are also widely available. Moreover, I rank them higher than academic programs. Why? Because they’re agile and provide up-to-date knowledge. As a rule, online courses are introduced by hands-on experts who share real-life experiences. Besides, online courses let you combine education with your other main activity, be it a full-time job, travelling, or anything else.
The drawback of online education is the lack of face-to-face communication. However, this should not be a problem nowadays due to the effect the COVID-19 has had on numerous industries. Anyway, it’s up to you.
Is math needed to master data analytics?
It’s highly recommended. Mathematics along with statistics would be a perfect aid to your education. For example, you’ll be able to differentiate between a median, an arithmetic average, and a mode. This will help you develop critical thinking skills.
In analytics, we don’t use complex methods, so it’s not about advanced mathematics. You can get this background at your bachelor degree course or find specialized courses on Coursera, Udacity, and other educational platforms.
Did you present any data analytics online courses yourself?
Yes, I’m a mentor in a data analyst course at Yandex Praktikum. Since I was a child, I’ve wanted to teach. I probably inherited this from my mother :). Anyway, I was looking for a way to share my experience in data analytics and encourage other data enthusiasts. This project allowed me to put this idea into effect.
What I do is run webinars dedicated to a specific topic based on real-life use cases. I can explain the intricacies of some approaches, share some code, etc. Sometimes I also check the projects that are submitted by students. The best thing here is it’s a win-win learning situation. I not only suggest my ideas on how to make their work better, but I also can learn from improvements offered by the students. This is very inspirational to see how the students’ skills develop from the beginning of the course.
What is the best time for a business to hire a data analyst?
In every business, you analyze data from the very beginning, when the business has not yet been launched. This process should never end, unless you want to leave the business. Analytical tasks can be distributed among various executives – project managers, marketing experts, etc. A data analyst is a person who can own this process. So, it’s better to have such an expert when the scope for analytics is large enough.
For example, you have a team of developers who are working on your product. Do you want them to spend their time on competitor analysis, communication with early adopters, and other analytical tasks? Perhaps it is allowable at an early stage, when the analytical scope is small. But later, you’ll need to hire a person who will own this process and focus only on data analytics.
How many data analysts should a mid-sized business have?
The number of data analysts you need depends on the scope and size of analytical tasks. First, you need to determine a business need: Will the work of a data analyst bear fruit for your business? The size of your business doesn’t matter; you may have a small company with 10+ data analysts in staff that will make the difference. For example, most brokerage firms rely on data analysis.
If you have the scope, hire one data analyst and see whether he/she can handle it. After that, you’ll better understand how many experts you need.
Data analytics at Coupler.io: What can a data analyst do on a product team?
Elvira participates in different projects by Railsware, including Coupler.io. It is a product that allows users to import data into Google Sheets from third-party data sources, such as Airtable, Trello, BigQuery, and more. Let’s find out her data analytics scope within this project.
What types of data do you analyze at Coupler.io?
- User activity
- How many rows users import
- How many data sources they import from
- Support requests
- Coupler.io blog performance
- Google Analytics
- Email activity
Examples of data analysis
We recently introduced a blog post, Post Coronavirus Destinations Dashboard, where you can see one of the examples of my scope.
And here is a competitor analysis dashboard we use to monitor our position on the G Suite Marketplace:
QUIZ: One day in the life of a data analyst
We asked Elvira to describe her scope through the example of a recent data analytics task.
A data analytics task
Create a table, which will show Coupler.io’s position in search results on the G Suite Marketplace.
- Google Cloud Functions (as an environment)
- Google Sheets
Around 6 hours
What do you like/dislike about data analytics?
- Constant learning and development.
- Data analytics is magic that we use to process data and make it getable and understandable.
- Data analysts are like detectives who need to investigate a case, find clues, and come up with an answer.
- Follow the trend – hire a data analyst. Some companies do not have scope for data analysts, but they still hire them. Why? It’s like a tribute to the modern trend: having a data analyst on board is up-market.
- Monkey business. A data analyst can do a good job that will be of no practical use. This is mostly associated with a lack of understanding of the data analytics scope.
- Data diddling. Usually, a data analyst makes an analysis and then comes up with a conclusion. But, what if the conclusion is already defined by a higher-level manager or investor? In this case, a business owner asks them to diddle the actual data analysis and show the metrics expected by a third party. For example, a data analyst sees that sales have not grown, but the investor is expecting them to have done. It has nothing to do with data analytics.
Afterword: How does data analysis help you in your day-to-day activities?
Data analytics helps me a lot in managing my finances and investments. My knowledge let me automate some data imports and create insightful reports to keep my eye on the ball.
I had a deep interest in feminism-oriented topics and analyzed the UN survey on violence against women in Kazakhstan. I also analyzed the OECD data on the gender wage gap. If you know Russian (or trust the power of Google Translate), check my findings on my Medium profile.
Country Chooser dashboard is another research I made to monitor the quality of life and climate data of the countries I want to visit in the future. Some of them I have already visited.
I hope my answers helped you understand why data analytics is so in demand. From my side, I’d like to wish you luck in this turbulent time. All the best!Back to Blog