Sourcing with YouTube

Google search: YouTube number of users

Result: 1.3 billion users

Google search: world population

Result: 7.5 billion

On YouTube, I searched for software engineer.

The first result was A Day in the Life of a Software Engineer.

Other results:

  • How I Spend My $250K Software Engineer Income
  • Technical interview with a Google engineer
  • Career Paths for Software Engineers and how to navigate it.
  • Why Software Engineering is hard
  • A Day In The Life of a Facebook Software Engineer Intern

Hold on. I must watch this last video…

OK, it was about nothing.

However, it had 254 comments.

Browse more results:

  • A Day In the Life of a Software Engineer Intern In San Francisco
  • What do I do as a Software Engineer?
  • A Day in the Life of a Software Engineer | New York City

Then, the famous:

  • How to: Work at Google — Example Coding/Engineering Interview

Then more of:

  • A Day in the Life of Software Engineer, Jakarta
  • 💻 One Day in Life of SAP Software Engineer
  • Day In The Life: Software Engineer Intern @ WeWork

On the recommendations, I saw this video from MIT Dynamic Programming I: Fibonacci, Shortest Paths. It had 568 comments.

This has to be valuable data…the comments.

Scraping Youtube comments

How do I scrape these?

Google search query: Scrape youtube comments

Found this Github page youtube-comment-scraper.

It’s built on nodejs. It has a website that says Enter Youtube URL.

Let’s try the MIT video I found.

Article Continues Below

It says “Scraping comments”…sip of coffee.

And it’s done.

Youtube comments

It gives the option to download a JSON or a CSV.

Let’s do CSV.

The file has this data:

  • id
  • user
  • date
  • timestamp
  • comment text
  • likes
  • hasReplies
  • numberOfReplies

A lot of interesting data in comments. Look at this comment:

What I'm missing in the code?
def fib (n):
 f=[]
 for k in range(1,n+1):
  if k<=2: 
   f=1
  else:
   f=f[k-1]+f[k-2]
  f[k]=f
 return f[n]

for i in range(100):
 print(fib(i))

The comment has the full name of the username. The date is from a year ago.

Googled the name. Found his Github. Now his email. Found him on Linkedin.

You could compose a message like this.

subject: fibonacci and shortest paths

Hey [don’t forget to put his name here],

I ran into a youtube video from MIT about the shortest path algorithm. I don’t want to sound like a creep, but I found your comment and wonder if anybody replied to your question. If they didn’t. Here is what might be missing from your code. [Add what the team suggested]. I know it has been a year, and I am sure you are beyond this. Anyhow, my team is looking for someone in [add city] to help us build [add what we are building]. We are currently [explain what we did]. I saw on your Github that you are working on [add project]. Super creepy I know 🙂

If you ever were to look for a new challenge, wondering what that would be. I am open to chat anytime. [add call to action]…

I love computers. My favorite tools are Linux, Python, Azure and AWS. I have two engineering degrees. Built teams at Sears. Now at WalmartLabs. Growing teams in Walmart eCommerce for machine learning, data science, data engineering and product management.

Topics