10 Good Coding Practices for Data Science

In this video, I talk about some good coding practices to be a strong programmer in data science. We talk all the time on this channel about the need to know at least one programming language like R or Python, but then once you're working with a language (any language), what are some of the best practices?

1) Define the purpose of the code
2) Minimize things that will (inevitably) need to be changed
3) Work from the top down, not to the right
4) Use descriptive variable and function names
5) Use functions rather than repeating code
6) But don't go overboard with chaining functions
7) Pay attention to datatypes
8) Provide adequate documentation and comments
9) Check for the latest and greatest packages for the task
10) Test as you go

#DataScience #BreakingIntoDataScience #ProgrammingForDataScience

BTC: 3LM5d1vibhp1F7pcxAFX8Ys1DM6XLUoNVL
ETH: 0x3CfC599C4c1040963B644780a0E62d45999bE9D8
LTC: MH8yPjvSmKvpmRRmufofjRB9hnRAFHfx32

[pipertripp]

My neck is sore from nodding with all of these points that you made here. I'm very new to data science and am only doing it as a hobby right now, but I've been coding for 27 years and all of these points are universally applicable and will make anybody a better coder.

[thebrutaltooth]

Awesome content as usual! I would love to see some functions that you have saved and use over and over again (in R please).

[Quienescribiohoy]

😂😂😂😂This Video is GREATE! It must be mandatory in any job induction! Keep the good work.. Your channel is awesome! Best regards form Argentina!

[fernandorosales]

Glad to have you back! Your videos are inspiring...let's go for the 50K subscribers mark 😎

[optimizacioneningenieria]

Great video Richard! It is nice you see you again.

[portillolopezjuanmanuel]

I'm glad you uploaded a new video, keep it up!

[bokai]

Before deploying or intergrating your model with other people's work:
Take a deep breath, expect problems, make some time to resolve unexpected issues

[alexwa]

I think it would be awesome to have some guidelinse/ roadmaps how you would plan a data engineer/ analyst career.
From getting your first job until the end of career. In my opinion it is real good to have a plan, also if you never have to stay strict at this.

[arcadevampire]

Great video, good points. Agree with 3, i hate code that runs off far to the right. I work on laptop screen and usually split it, so try to code as pithy as possible on left side, so my code thin and long. Is that bad too? What's worse wide code or long code?

[pipertripp]

One that you didn't mention... using vectorized calls instead of writing loops where ever possible to improve performance. Folks that are new to R and Python might not realise how great this concept is. It look me a little time to get my mind round it when I first encountered it with numpy. I think numpy nicked it from matlab. There's nothing new under the sun I suppose.

[tourettesdisorder]

Python or R? Hammer or screwdriver? Who knows?

[xaviercasas]

best practice comment steps so i don't get lost