Data Analyst Career Guide: Role, Salary, Skills, and Career Paths

Common job titles include Data Analyst, Business Data Analyst, BI Analyst, Product Analyst, Marketing Analyst, Operations Analyst, and Growth Analyst. Although the titles differ, their core function is to use data to answer business questions.

For example, if an e-commerce company's monthly sales decline by 15%, the analysis would not stop at the result of "a 15% decline." Instead, sales would be broken down into the number of users, conversion rate, and average order value, and then the source of the change would be investigated by channel, region, device, and user type.

• The number of users is basically stable, indicating that the problem may not lie in traffic.
• The order conversion rate has clearly declined, requiring further breakdown.
• The conversion rate drop is highest among Android users.
• Upon further investigation, it was found that the payment page in the new version of the app had issues.

The final conclusion is: the main reason for the decline in sales is not decreased traffic, but that the new Android payment process affected conversions. This process—from phenomenon to cause, and then to solution—is what constitutes complete data analysis.

① Turn vague requirements into computable questions

When the business asks, "Has user churn recently worsened?" the analyst first needs to clarify the observation period, user scope, and churn definition, and then decide whether to look at 7-day retention, 30-day retention, overall churn rate, and the changes across different channels, regions, and user groups.

② Acquire data from the database

Most companies store their business data in databases, so SQL is one of the most important skills for data analysts. Analysts need to select fields, join tables, restrict time ranges, and aggregate results based on the questions.

③ Clean and verify data

Real-world data often have issues such as duplicate orders, missing values, incorrect date formats, abnormal records, and inconsistent field names. Cleaning is usually done through Excel, SQL, or Python Pandas, while also verifying metric definitions to avoid different departments using the same name but calculating different results.

④ Analyze data

Common methods include year-over-year, month-over-month, averages, medians, conversion rates, retention rates, user segmentation, funnel analysis, correlation analysis, A/B testing, and trend analysis. The methods themselves are not difficult; the challenge lies in choosing the method that can truly answer business questions.

⑤ Visualize and build dashboards

Results are usually presented as bar charts, line charts, funnel charts, maps, or dashboards. Common tools include Power BI, Tableau, Looker, and Excel. The goal is not to make the charts fancy, but to allow readers to quickly understand the problem.

⑥ Provide conclusions and recommendations

“Orders in July decreased by 12%” is only a description of the data. A more complete conclusion should point out: the order decline mainly comes from new users, whose orders dropped by 28%, with 70% of the drop coming from a specific channel; the customer acquisition cost for this channel increased while the conversion rate for new users decreased, thus the placement strategy needs to be reassessed.

The complete chain is: data → information → cause → conclusion → decision.

Overall prospects are still good, but positions are becoming clearly differentiated. Simple tasks such as creating summary tables, repeatedly preparing daily reports, and making charts according to fixed templates are increasingly being replaced by automated SQL, intelligent BI, and generative AI tools.

In the 2026 career classification, business intelligence analysts are still listed as positions with good prospects. As a more technical adjacent occupation, data scientists are expected to see employment growth of about 34% from 2024 to 2034, significantly higher than the average of about 3% for all occupations.

The skills with long-term value are shifting to:

• Understanding business goals and profit logic.
• Designing reasonable indicator systems.
• Conducting user, product, channel, and experimental analysis.
• Establishing stable data models and analysis processes.
• Using AI to improve the efficiency of data collection, cleaning, analysis, and presentation.
• Translating analytical conclusions into actual decisions.

Data analysts will not disappear because of AI, but positions that only involve "pulling data and making reports" will continue to decline in value. Those who can define problems, determine calipers, identify causal boundaries, and drive business actions will become even more important.

A bachelor's degree has the most advantage in recruitment. Majors such as statistics, mathematics, computer science, economics, finance, information management, business analysis, and marketing are highly relevant to the position. However, major is not the only requirement; people from accounting, operations, marketing, supply chain, sales, and engineering have also successfully transitioned into data analysis.

Business intelligence analysis is usually classified as a position that requires relatively adequate professional preparation. Most related jobs tend to require a four-year bachelor's degree, but not all positions set the degree as an absolute requirement. Sample open recruitment in China for 2026 also shows that about 65.5% of junior data analyst positions require a bachelor's degree, which is also the largest educational level among internship positions.

Can an associate degree work?

Yes, but the difficulty is usually higher during the screening stage. It is necessary to make up for the educational gap with a portfolio, SQL skills, real projects, and industry experience. Rather than accumulating course certificates, it is better to present projects that can demonstrate the complete analysis process.

Is a master's degree necessary?

Ordinary data analyst positions usually do not require a master's degree. If the goal is data science, quantitative finance, algorithms, AI, or core data positions in large tech companies, graduate-level training in mathematics, statistics, and research will be more valuable.

📗 Excel

• VLOOKUP, XLOOKUP, SUMIFS, COUNTIFS, and IF.
• Pivot tables, data cleaning, conditional formatting, and common charts.

🗄️ SQL

• SELECT, WHERE, GROUP BY, ORDER BY, and JOIN.
• CASE WHEN, HAVING, subqueries, CTE, and window functions.
• The level of being able to independently design queries and extract data after seeing a business problem.

📐 Statistics

• Mean, median, variance, standard deviation, probability, and normal distribution.
• Sampling, confidence intervals, hypothesis testing, correlation, regression, and A/B testing.
• Understand the applicable boundaries of statistical conclusions and do not directly interpret correlation as causation.

In addition to mathematical foundations, critical thinking, problem-solving, judgment and decision-making, and communication skills are also important abilities for this type of position.

📊 BI Tools

Master at least one of Power BI or Tableau, able to establish metrics, design filter logic, and create maintainable dashboards.

🐍 Python

Focus on learning Python basics, Pandas, NumPy, Matplotlib, data cleaning, and data analysis. Python is not a strict requirement for all junior positions, but it helps handle larger data sets, automate repetitive tasks, and broaden long-term development opportunities.

💼 Business Skills

E-commerce requires understanding GMV, average order value, conversion rate, repurchase rate, CAC, LTV, and ROI; internet products require understanding DAU, MAU, retention rate, churn rate, and activity rate; advertising business requires understanding CTR, CPC, CPA, and ROAS. Being proficient in SQL but unable to understand the business makes it difficult to become an outstanding data analyst.

People with experience usually find suitable positions more easily, whereas job seekers with zero experience face stronger competition. Data analysis positions are not the most numerous in most companies; a company may have a large number of sales, customer service, and operations staff but only a few analysts.

Graduates in statistics, computer science, economics, and business, as well as professionals wishing to switch careers from operations, finance, and marketing, will all compete. Therefore, learning Excel and SQL does not mean you can immediately get a job. Companies truly care about whether you can understand the business, complete analysis independently, and provide reliable conclusions.

Looking only at the entry level, the technical threshold for data analysis is lower than algorithm engineering, machine learning, and software development. Junior positions mainly use Excel, SQL, BI tools, and basic statistical knowledge.

The real difficulty is "analysis" itself. Facing questions like "Why has the number of users decreased?", you need to determine which metrics to look at, how to break down the data, how to validate hypotheses, and whether the decline comes from traffic, conversion, specific channels, or a certain type of user. This ability depends on long-term accumulated business understanding, logical skills, and practical experience.

The most realistic path is: Basic tools → Analysis methods → Project practice → Job preparation.

• Month 1: Excel and basic data analysis.
• Month 2: SQL queries, joins, aggregation, and window functions.
• Month 3: Power BI and basic statistics.
• Month 4: Python data processing.
• Month 5: Complete projects and organize a portfolio.
• Month 6: Prepare your resume, apply for positions, and practice interviews.

If you study steadily for 10–15 hours per week, it usually takes 4–6 months to reach an interview level for entry-level positions. This does not mean you will definitely find a job within 4 months. For those completely changing industries, without an advantageous degree or experience, the preparation period may reach 6–12 months.

The most common problem for beginners is taking a lot of courses without completing a project they can showcase. A more effective way is to learn and do at the same time:

• After learning Excel, analyze a set of e-commerce sales data.
• After learning SQL, create an orders database and calculate the number of users, sales, repurchase rate, average order value, and popular products.
• After learning Power BI, create a set of filterable sales dashboards.
• After learning Python, complete a user behavior or retention analysis.

Before job hunting, it is best to prepare at least 3 complete projects, such as e-commerce sales analysis, user retention analysis, and marketing campaign analysis. Each project should include the problem background, data description, cleaning process, metric definitions, analysis methods, key findings, and business recommendations. Three complete projects are usually more persuasive than ten course certificates.

SQL

Common questions include querying the top 10 products by sales, extracting each user’s most recent order, identifying consecutive login users, calculating 7-day retention rate, and using window functions for ranking or cumulative calculations.

Excel

Common issues include the difference between VLOOKUP and XLOOKUP, using pivot tables, handling duplicate data, and multi-condition aggregation.

Statistics

Key points include the applicable scenarios for mean and median, the meaning of statistical significance, basic design of A/B testing, and why correlation does not equal causation.

Business cases

Facing a situation like ‘company sales suddenly dropped by 20%’, don’t directly guess the reason. You can first break down sales into order count and average order value, then break down order count into user count and conversion rate, and gradually narrow down the scope from dimensions like new vs. returning users, channels, regions, and devices.

Project experience

Interviewers usually follow up on project goals, data sources, metric selection, cleaning methods, analysis process, main findings, and final recommendations. Every number and conclusion on the resume needs to be explainable.

A self-introduction should ideally be controlled within 60–90 seconds, covering background, skills, projects, results, and reasons for applying.

A clear way to express this could be: My undergraduate major was Marketing. Over the past year, I have mainly studied and practiced data analysis, and I am now familiar with Excel, SQL, Power BI, and Python. I have used SQL to process around 200,000 order records to complete analyses of user retention, repurchase, and consumption behaviors, and created sales dashboards with Power BI. The analysis revealed that high-value customers, though only about 12% of the total user base, contributed close to 40% of sales, leading to a membership marketing plan targeted at high-value users. I hope to continue developing in the field of business data analysis.

The emphasis is not on generic descriptions like 'cheerful personality, strong learning ability' but on proving through specific projects that you can indeed perform data analysis work.

The job transition range for data analysts is very broad. Common directions include Product Manager, Business Analyst, BI Engineer, Data Scientist, Data Engineer, Growth Analyst, Market Analyst, Financial Analyst, Strategy Analyst, Operations Manager, User Research, and Consulting.

• Transition to Product Manager: Suitable for those who are already familiar with users, products, data, and metrics systems.
• Transition to Data Science: Suitable for those with strong technical interests and a willingness to continue learning mathematics, statistics, and machine learning.
• Transition to Business Analysis: Suitable for those with strong business understanding, expression, and communication skills.
• Transition to Data Engineering: Requires supplementing skills in databases, data warehouses, ETL, cloud platforms, and engineering development.

• 09:00: Check core metrics such as yesterday's DAU, sales, order volume, and conversion rate to see if there are any anomalies.
• 10:00: Receive business requests, for example, analyzing the reasons for a recent decline in new user repurchase rates.
• 10:30: Write SQL to extract registration, access, purchase, and repurchase data.
• 13:00: Clean data, check for missing values, duplicate records, outliers, and statistical standards.
• 14:00: Conduct user segmentation analysis by channel, region, device, age, and time.
• 15:30: Discover a decline in the quality of new users from a certain channel and further verify the reasons.
• 16:00: Confirm with the marketing department whether recent placement strategies have changed.
• 17:00: Create charts or update dashboards.
• 17:30: Organize problems, causes, impact scope, and action suggestions.

Actual work usually simultaneously involves data processing, report maintenance, dashboard updates, exception investigation, and cross-department communication, and does not involve doing complex models every day. Job task descriptions also emphasize analyzing data, processing information, generating reports, and explaining results to managers and other departments.

Data analysis is usually a medium-pressure position. It does not bear performance indicators directly like sales, nor does it need to continuously handle large numbers of clients like customer service, but it faces significant pressure regarding data accuracy.

In related occupational surveys, 64% of respondents consider work accuracy 'extremely important.' Once numbers are included in business reports or management decisions, query errors and standard deviations can have real impacts.

If you report a 10% profit decline to management and later find that the SQL was written incorrectly or the standard was used wrongly, the impact can be very large. Therefore, reviewing query logic, conducting sample checks of data, confirming standards with the business, and retaining calculation processes are all important.

Temporary requests, report follow-ups, end-of-month and quarterly summaries, event reviews, and data anomalies can also bring pressure. The level of pressure is usually directly related to the company's pace, team configuration, data infrastructure, and management style.

The need for overtime is mainly determined by the industry and company culture, not by the job title.

• Traditional enterprises: Work pace is usually relatively stable.
• Internet companies: Temporary requests are frequent, and some teams often work overtime.
• E-commerce: Busy during major promotions, events, and reviews.
• Gaming: Workload increases during product launches, version updates, and events.
• Finance: Some core positions have a fast pace and high accuracy requirements.
• Consulting: Project delivery periods can be very busy.

In related career surveys, 91% of business intelligence analysts consider their work schedules to be regular and fixed. Data analysis is not inherently a profession that requires long-term high-intensity overtime, but if the company overall operates long hours, analysis positions usually cannot be an exception. Therefore, when job hunting, it's necessary to understand the team's pace, demand management methods, report automation level, and overtime compensation system.