Still looking for a New Year’s resolution? Here are 6 skills to develop to make you & your team ridiculously efficient.
What makes a great data analyst? Great data analysts can find creative solutions to complex problems and produce quality work in record time.
They know exactly which questions to ask to get to a robust problem statement; from there, they know exactly which process to follow, which query to write, which dataset to use, and how to make the insights as digestible as possible.
They make it all look so easy… but what is their secret?
In short — they have developed the right set of skills. They trained hard to develop the right muscles, making them a rich blend of different capabilities. Let’s dive into their gym schedule — spoiler alert, they didn’t skip “stats” day.
SQL is the language of data analysis. It is critical to be fluent in it to be able to delve and derive deeper insights. And by fluency, I don’t mean proficiency — I truly mean fluency, i.e. not thinking twice before putting together a 100-line script with multiple CTEs, using arrays and window functions.
The lack of fluency in SQL can greatly limit an analyst. Either they become dependent on others for data retrieval — which greatly limits their execution speed — or if they are only relying on their skills, they become forced to stay at the “surface” of insights, potentially missing the deeper, valuable truths beneath.
For an analyst to achieve fluency, there are no secrets:
- Practicing regularly: having daily/weekly challenging sessions, working on complicated projects pushing them outside their comfort zones
- Learning from others: reviewing the code of more knowledgeable colleagues, participating in internal/external online forums, and/or taking structured courses
Statistics is scary for a lot of people and for a good reason — it can quickly become very complex. At the same time, having a solid grasp on a few key concepts can generate a ton of value, and allow to find creative ways to answer not-so-easy questions.
Most of the great analysts I worked with had the following:
- A solid grasp of descriptive statistics. Arguably, this is crucial for any descriptive or exploratory analysis, and it sets the stage for more complex analyses
- A good understanding of the difference between a population and a sample, how that relates to statistical testing, and how to do some common statistical tests
- Bonus point: a rough understanding of machine learning: what are some of the key principles, how to evaluate the performance of a model, etc.
When working with data, it is easy to feel like “you understand” the machinery. You know the numbers. You know the trends. But without the domain knowledge, i.e. without the qualitative side, it is easy to miss some key insights. Because at the end of the day, a dataset is just a simplification. It offers a limited and simplified lens to look at a phenomenon. Domain knowledge is what gives the additional context needed to understand what cannot be seen in the dataset itself.
While it is possible to acquire domain knowledge “just” by staying in the same company/industry for years — it is possible to ramp up faster by being intentful about it. Great analysts usually do a mix of those 3 activities:
- They shadow colleagues: they make friends with their cross-functional partners and actively try to understand their day-to-day job
- They regularly discuss their quantitative findings with subject matter experts to incorporate qualitative insights and validate their data interpretations.
- They read industry reports, they follow “Linkedin Influencers”, they participate in industry-specific events, discussions, etc.
A lot of time is usually spent in finding the right data source (or logic) to use for a given project. One of the reasons why great analysts are quite fast is because they have developed a large knowledge of the different data sources available, including their specificities… and their oddities. They directly know where to find the information necessary for their project, and which actual logic to use — because they know how the data is transformed, and where it is being housed. To achieve this:
- They are curious about the data journey: they mapped how the data ended up in their favorite dataset back to raw data and gained a clear picture of its lifecycle and potential points of quality degradation or enhancement.
- They collaborate with data engineers: they discuss with them regularly; they don’t hesitate to reach out every time they face a new “oddity”; they try to understand their challenges and objectives to make sure they align their analytical work with the technical realities of the infrastructure.
Every company uses different tools, and each tool has different capabilities and limits. A lot of analytical tools have hundreds of pages of documentation so it is easy to miss out on some of the great capabilities they can have. But great expertise of the tool can be a game changer — and great analysts have understood that:
- They explore the advanced features of the tools that are given to them — through tutorials, by reading forums, and by simply practicing
- They test how they can integrate the different tools with each other and try to automate the most repetitive tasks to free up time for deeper analysis.
- They try to stay updated (by joining online communities — e.g. Reddit) and to continue experimenting with innovation in data tooling
Last but not least, as an analyst, having good business acumen can help you understand which insights are more valuable; how to make those insights more digestible to your audience; and how to make sure your organization will derive as much value as possible from your studies. There are a few ways great analysts go about sharpening their business acumen:
So, what’s the secret sauce that makes a great data analyst? It’s about building a robust skill set. It’s about a holistic development of skills. These analysts don’t just rely on one aspect of their expertise; they develop a harmonious blend of technical, statistical, and business acumen.
In summary — It’s about not skipping “stat day” — or any of the other muscle days. Just like at the gym.
