The “Analyst” title is widely used across different industries and functions (just think of finance analysts, marketing analysts, or sales operations analysts). In this paper, we will focus on data analysts- those analysts who have a few more tech skills (advanced Excel, SQL, and sometimes more), and are specialised in getting the right data to the right place and in the right format.
Just within the data world - this can cover a wide range of titles and job descriptions - not all of which will be covered in this post. For example - there are data analysts, business intelligence analysts, data warehouse engineers, analytics engineers, product analysts, growth data analysts, and more. This paper will focus on those analysts having – and using on a weekly basis – SQL and dashboarding/spreadsheets as their core skill set, namely data analysts, business intelligence analysts, and sometimes product analysts. We will dive further into the other roles more in future!
According to LinkedIn - there are 820,000+ data analysts professionally engaged on their platform, and approximately 400,000 open data analyst roles worldwide.
When we focus on data/BI/product, this does not mean these positions are mutually exclusive from finance or marketing or sales analysts - the skill set can sometimes overlap. The main difference is usually a specific domain function - a marketing analyst might have passed a Google Analytics certification, and a finance analyst might have a degree in accounting. For data analysts - one common prerequisite on job descriptions is a strong numerical training - e.g. Mathematics, Econometrics, Statistics, or Computer Science. As much as the skill sets overlap - a data analyst will be responsible for the more complex queries, data work and spreadsheet analyses, while other analysts would usually get their work done with simple SQL queries or sometimes no SQL, and maybe doing the complex analysis within a spreadsheet (e.g. finance analysts).
According to an Indeed analysis - the top three skills for a data analyst are SQL, spreadsheets/dashboarding, and critical thinking. SQL is a core technical skill - being able to write the select statements that pull data. Spreadsheets and dashboarding are important to get insights from the data, after it has been extracted using SQL. Critical thinking is important throughout - to think through the business challenge and turn it into a data question.
Structured Query Language (SQL) is the bridge between the data output of data engineers and the data points that business stakeholders need to understand their business, understand performance, monitor trends and KPIs, and create reporting.
A data analyst will spend a significant amount of their time writing SQL. Often it starts with exploring datasets. For example - the first few weeks at a job, an analyst might explore some random rows of data from various tables to see what they look like, the type of data they contain, and figure out how different tables will connect to each other. This is critical knowledge to develop for the future - so they know the full breadth of data that can be leveraged to solve a business challenge. A good analyst is often like a good librarian in this case- if an avid reader comes in off the street and asks for a book, a librarian will be able to answer straight-away if it is available or not, and when discussing business challenges with business stakeholders, a great data analyst can speak knowledgeably as to what data is available (and what is not) to help solve the problem and get to an action plan quicker.
Another major use of SQL is writing the queries that will go into scaled tools - such as dashboards or other types of company reporting. These queries are usually a bit more complex in that they require a high degree of standardization and optimization - agreement on definitions of metrics and dimensions to be included, as well as complex inner queries to get the report aligned with company definitions.
Once an analyst is ramped up and comfortable with a company’s data, a significant amount of time is spent writing SQL queries to deep-dive into one-off questions from the business - what happened to revenue today? Why did costs go up last month? What happened to our top Client’s performance? These are often questions which arise from standardized reports that a data analyst has previously built - a manager might see that revenue has increased 10% in one country or vertical, and ask the data analyst to deep-dive further - did this come from a particular city? Or a particular salesman? This usually requires small changes to an existing SQL query (e.g. the one used in a dashboard or company report), good knowledge of the datatables to know what dimensions or metrics can be added for the deep-dive analysis, and usually lots of patience as the results are returned.
Spreadsheets are an equally important part of a data analyst’s technical expertise, and often a preliminary step towards a dashboard. A data analyst can sometimes copy-paste data or download a csv for a business stakeholder to get an initial look at the data, but often goes further to build a more comprehensive dataset in a spreadsheet - using VLOOKUPs, or other functions. The next step is a pivot table. After adding any relevant groupings to the dimension, or appropriate calculations to the metrics- the right calculated field, or the right aggregation (avg vs sum), as well as formatting the table to make it visually appealing and easy to read, this pivot table can become the first level of reporting. Graphs and charts can also be added to make the data more digestible. Oftentimes - the above steps can be bypassed by connecting a dataset directly to a dashboarding tool - such as Looker, Tableau, or Data Studio.
Then - the data analyst is responsible for the accuracy and standardization of the dataset, and business stakeholders can build out dashboards to their liking. The recent trend has been for data analysts to get more technical and focus more on getting the right data to the visualization tool - and from there, the dashboard tools can expedite the work previously done inside spreadsheets.
SQL and spreadsheets are standard tools in companies using big data- from here the tools and technical skills start to vary. Some companies might use Looker, or Tableau or Google Data Studio and ask for some dashboarding experience in their data analysts. Another increasingly sought-for skill is Python. Python deserves special mention as it is slowly but surely becoming a prerequisite for most analytical roles (not just data analyst). It became the most popular programming language in March 2021, and is showing up on more and more job descriptions. Also - with the power it gives data analysts to more quickly and directly access multiple data sources and connect them to spreadsheets - it is becoming a key skill for data analysts to do their work more quickly and efficiently.
The third most-sought skill in data analysts is critical thinking. While this sounds like a cliché, it is especially important for data analysts so they can do their work with minimal supervision and hand-holding. The questions they are asked - what happened to revenue yesterday? What happened to our top Client? - are often open-ended, urgent and expect thorough answers with minimal hand-holding. Data analysts’ academic studies in the usual fields of Mathematics, Econometrics, Statistics, etc help train their critical thinking and ability to think through the problem and identify the key elements needed for a great answer. This critical thinking is necessary to look at a business challenge in a logical and structured manner, and great to pair alongside an extensive understanding of the data available.
After having spoken about the breadth of expertises necessary for a data analyst- it also serves to speak about the depth of each of these expertises- how much time, on average, a data analyst could expect to find him/herself working on each of the above tasks.
A common breakout usually looks like ~40% SQL, ~40% dashboarding/scaling/presentations and ~20% business/stakeholder projects, but this will also vary greatly by company and the stakeholders one is working with. At some extremes (e.g. large tech companies), the dashboarding can be replaced entirely by more localized spreadsheets, albeit still powered through SQL and possibly requiring less time. In some roles which support sales more closely - the primary need is to pull filtered client data and then make great sales presentations.
Any down time is usually spent at the ping-pong table in heated competition with stakeholders, or discussing a business challenge over a latte from the coffee bar.
This is complex because there is no universal definition of data analyst - this can vary drastically across companies and industry. These roles can be very complex, and work with a wider variety of tools- One of the first tasks is deciding the core requirements, key skills and tools used in your company, and then tailoring an assessment to find the right candidate.
Skillfill is a unique, novel tool that generates skill tests for your applicants to take. All you have to do is upload your job ad. The Skillfill AI-engine then extracts all required skills and translates them into a highly specific skill test. The hiring manager can then see the applicant’s results and decide which are the most interesting candidates to move forward with. The test results are broken out by skill category and also, by individual question, so the hiring manager can get a better understanding of the applicant’s strengths and weaknesses, and compare them to the team’s hiring needs. It is the tool that enables HR to assess their candidates with the expertise level of your data department without involving them in an assessment creation. The usage of the Skillfill tool ensures that HR gets the most suitable candidates in front of the hiring manager.