Karina Data Scientist

Data Science & Analytics Expert.
I like to share tips and tricks

[Karina Data Scientist @KarinaData]

2/2 sketchbook style, vintage magazine illustration on beige aged paper. Add handwritten notes, arrows, funny observations, doodles, and absurd personality analysis around the character.

[Karina Data Scientist @KarinaData]

Turn photo into a sketch with ChatGPT.

Prompt below:

[Karina Data Scientist @KarinaData]

1/2 Turn the person from the photo into a grotesque humorous fashion caricature sketch. Strongly exaggerated anatomy, elongated proportions, expressive crooked hands, dramatic face, awkward elegant posture, satirical fashion energy, messy ink lines, rough watercolor,

[Karina Data Scientist @KarinaData]

Most SQL problems analysts solve with subqueries can be solved in one line.

Window functions do that.

Here is the simplest way to understand them

A window function performs calculations across rows…

WITHOUT collapsing the dataset like GROUP BY.

[Karina Data Scientist @KarinaData]

AI certifications are not hard to get. Finding ones that actually mean something is.

No courses that cost thousands. No badges from platforms nobody has heard of. Just 8 credentials that hold weight.

[Karina Data Scientist @KarinaData]

Python statistics cheatsheet

[Karina Data Scientist @KarinaData]

Advent of Code — https://t.co/LuBnVbYCRK
Making Games with Python & Pygame — https://t.co/eoNiyX6Zp6

[Karina Data Scientist @KarinaData]

You can learn Python by playing a game.
Codédex — https://t.co/3IQAjKRaQR
CodinGame — https://t.co/5Mzz48y0pm
Codewars — https://t.co/Umz90HJwzy
Exercism — https://t.co/ZiVkydPZLy
Battlesnake — https://t.co/Af6VWZO0td
CheckiO — https://t.co/dRT653FC0J

[Karina Data Scientist @KarinaData]

If you work with data in Python, these are the functions you will use on repeat.

That last one. A left join and an inner join look similar.
The results are not. Caught me out more than once.

[Karina Data Scientist @KarinaData]

I’ve never worked with a clean dataset.

Every real project = messy data.

And it always comes down to 4 things:
• Missing values
• Duplicates
• Data types & formatting
• Outliers

Cleaning isn’t a “prep step”. It is the analysis.

[Karina Data Scientist @KarinaData]

GitHub is one of the most underused free resources in data.

Here are 10 repositories worth bookmarking, covering data analysis and data science:

[Karina Data Scientist @KarinaData]

Customers in this month's cohort but not last month's (and vice versa)

One symbol. No loop. No if statements.

[Karina Data Scientist @KarinaData]

Python set operators analysts actually use

You already know sets remove duplicates. But they also do something more useful.

Compare lists without a single loop.
| union -- combine two lists, no duplicates

All customers who bought in January OR February

[Karina Data Scientist @KarinaData]

& intersection -- what appears in both

Customers who churned AND never contacted support

- difference -- what's in one but not the other

Products sold in UAE but not in KSA

^ symmetric difference -- what's exclusive to each side