Karina Samsonova

Karina Samsonova

These are the best posts from Karina Samsonova.

4 viral posts with 2,347 likes, 307 comments, and 177 shares.
0 image posts, 0 carousel posts, 0 video posts, 0 text posts.

πŸ‘‰ Go deeper on Karina Samsonova's LinkedIn with the ContentIn Chrome extension πŸ‘ˆ

Best Posts by Karina Samsonova on LinkedIn

Unpopular opinion: live coding interviews for data analyst roles are a waste of everyone's time.

I have held dozens of interviews over the years and been through plenty myself. I have never once done a live coding test. Here is why.

When I am stuck on a SQL function at work, I Google it. When I need to check syntax, I ask AI. When I am completely stuck, I ask a colleague. That is not a weakness. That is the job.

A live coding test measures how well someone performs under artificial pressure with no access to the tools they use every day. It does not measure how good an analyst they actually are.

What I tested instead:

Can you think through a business problem?

Give me a messy scenario and walk me through how you would approach it. No code required. Just thinking out loud.

Do you understand the logic, not just the syntax? I can teach syntax. I cannot teach someone to think analytically.

Can you communicate your findings? Most analyst work ends in a meeting or a slide deck. If you cannot explain what you found, the analysis does not matter.

Are you someone the team can work with? Attitude, curiosity, and honesty matter more than whether you can write a window function from memory on a shared screen.

When I interviewed for my current role I was allowed to use any tools I needed. That told me something about the company before I even started.

The best analysts I have hired were not the ones who aced a syntax test. They were the ones who asked the best questions.

What is your experience β€” live coding test or no?
Every data analyst CV I review has the same two fatal flaws.

No numbers. No tools.

Here's what to fix, section by section.

Professional summary

Two to three sentences. What you do, what tools you use, where you're headed. Career changers: frame your past as an asset, not an apology. Six years in finance is not irrelevant. It means you already understand the business questions behind the data.

Technical skills

List every tool you actually use. SQL, Python, Power BI, Tableau, Excel. ATS systems filter for these before a human ever sees your CV. If it's not listed, you don't exist in that search.

Work experience This is where most CVs fall apart.

Don't write: "Responsible for sales data analysis using Excel."
Write: "Analysed 50k+ monthly transactions in Excel to identify revenue leakage across 12 product categories, presented to senior leadership."

Every bullet needs three things: what you did, what tools you used, and what the result was.

Portfolio projects

Two or three projects with a one-line description, tools used, and a GitHub or Tableau Public link. If you are early in your career or changing fields, this section can outweigh years of unrelated experience.

The CV won't get you the job. It'll get you the interview.

𝐏.𝐒. I share my experience, data analytics & data science tips in my free newsletter. Join 3000+ readers here -> https://lnkd.in/d3M49ktj
Machine learning is not hard. The way you’re learning it is.

Here are 8 free resources.

1. Elements of AI β€” elementsofai.com

Built by the University of Helsinki. No maths, no coding required. Explains what ML is, how it works, and what it cannot do β€” in plain language. Over 2 million people have taken it. Start here if you have never touched ML before.

2. Google Machine Learning Crash Course β€” https://lnkd.in/dicK4Aia

Completely free, no account required, no paywall. 25+ lessons covering core ML concepts with interactive exercises. Built for Google's own engineers and made public. Requires basic algebra but nothing beyond that.

3. Practical Deep Learning for Coders β€” course.fast.ai

Designed for people with some coding experience who want to apply deep learning to practical problems β€” no advanced maths background required. Taught by Jeremy Howard, a former top-ranked Kaggle competitor. Completely free.

4. Kaggle Learn: Intro to ML β€” https://lnkd.in/difj7rC9

Free, in-browser, no setup required. Teaches ML through real datasets with immediate hands-on practice. One of the fastest ways to go from zero to running your first model.

5. Andrew Ng: AI for Everyone β€” https://lnkd.in/dXpqatRM

Free to audit on Coursera. No technical background required. Covers what ML is, how it affects business, and how to work alongside data science teams. Built for non-engineers who want to understand ML conceptually.

6. Microsoft ML for Beginners β€” https://lnkd.in/dVzMV-m9

A free 12-week curriculum on GitHub covering classic ML with scikit-learn. 26 lessons, quizzes, and projects. Assumes high school maths β€” nothing more.

7. StatQuest with Josh Starmer β€” youtube.com/@statquest

Free YouTube channel that explains ML algorithms visually and in plain English. If you have ever been confused by decision trees, random forests, or gradient boosting β€” watch StatQuest first. Over 1 million subscribers.

8. Hands-On Machine Learning with Scikit-Learn (free online) β€” https://lnkd.in/dzJGRYqQ

The full book available free on GitHub. Covers ML and deep learning with Python code throughout. Best used after you have the basics from one of the above resources and want to go deeper.

Save this for when someone tells you ML requires a maths degree.

𝐏.𝐒. I share my experience, data analytics & data science tips in my free newsletter. Join 4000+ readers here -> https://lnkd.in/d3M49ktj
I have read a lot of books on data, statistics, and career development. These 8 do not appear on most recommended lists β€” and they should.

TECHNICAL

1. Head First SQL β€” Lynn Beighley
More visual and beginner-friendly than most SQL books. Uses diagrams to explain concepts that trip people up. Good if traditional SQL books felt dry.

2. Python Crash Course β€” Eric Matthes
Learn by building real projects from early on. Less reference book, more learn-by-doing. The best starting point for complete beginners.

STATISTICS & THINKING

3. The Art of Statistics β€” David Spiegelhalter
Real cases β€” crime statistics, cancer screening, election polling β€” used to teach statistical thinking. Readable without being dumbed down.

4. Fooled by Randomness β€” Nassim Taleb
How humans confuse luck with skill. Directly relevant to any analyst interpreting results and presenting conclusions to stakeholders.

COMMUNICATION

5. The Visual Display of Quantitative Information β€” Edward Tufte
The original data visualisation bible. Where most modern visualisation advice traces back to. Dense but essential.

6. Made to Stick β€” Chip Heath and Dan Heath
Why some ideas survive and others die. Directly applicable to presenting findings in a way stakeholders actually remember.

CAREER & THINKING

7. Range β€” David Epstein
Why generalists outperform specialists. If you came from finance, audit, or another field before moving into data β€” this book will reframe that as an advantage.

8. The Lean Startup β€” Eric Ries
Not a data book. Required reading for any analyst working in a product or tech company. How data-driven decisions get made at speed.

𝐏.𝐒. I share my experience, data analytics & data science tips in my free newsletter. Join 4000+ readers here -> https://lnkd.in/d3M49ktj

Related Influencers