All micro-internships
Micro-internship brief ≈7h in one sittinglow-code

Audit a YouTube channel: what actually drives its views?

Creators guess; analysts check. Build a dataset of one channel’s last 30 videos — views, length, title style, upload gap — then use AI to find what actually correlates with performance, and verify it before you believe it.

Data AnalysisResearch & Fact-checkingSpreadsheet AnalysisData Visualisation

Posted by The AI Internships

The work brief

  1. 01Pick a mid-size channel you watch (not a megastar — patterns are clearer). Collect its last 30 videos into a CSV: title, views, duration, upload date, and features you code yourself (question title? number in title? thumbnail has face?).
  2. 02Normalise for age: views-per-day since upload beats raw views.
  3. 03Hand the CSV to AI and ask which features correlate with views-per-day — then verify its top claims against the raw rows and dig into the outliers yourself.
  4. 04Write the memo you’d send the creator: 3 evidence-backed recommendations.

What you’ll produce

4 deliverables

Submission standard

Submit the channel, the 30-video CSV with your coded features, the analysis, and a chart. The feature-coding is your real work — AI can’t watch the videos for you. Include your AI workflow — tools, best prompts, and what you changed. Add the channel's public link and a dated screenshot of its videos list, and make sure your CSV's views and dates match it.

  • The channel and why you picked it

    Short answerRequired

  • The 30-video dataset as CSV, including your hand-coded features

    30 videos including a video-URL column and your hand-coded features; views and upload dates must match the public channel on your collection day.

    CSV dataRequired

  • What correlates with views-per-day, the outliers explained, and 3 recommendations for the creator

    Written responseRequired

  • A chart of your strongest pattern

    Image uploadRequired

You’ll complete these inside your private workspace.

What you must submit as proof

This brief requires evidence an AI can’t fabricate.

  • Public link to the channel's videos page

    Lets a grader open the channel and spot-check your view counts and upload dates against the CSV.

    Public linkRequired

  • Dated screenshot of the channel's videos list showing view counts and upload dates

    Must show the real YouTube page with views/dates that match your CSV rows on the day you collected.

    Image uploadRequired

Submissions without this evidence cannot be submitted.

Protect other people in your proof. Blur faces, names, phone numbers and email addresses before you upload, and refer to anyone you worked with by role or number ("Listener 1", "the stall owner"). Your proof is only ever used to check your work — it is never published, never appears on your certificate, and is never shown in your public portfolio.

How your work is evaluated

The passing benchmark is 70/100.

Dataset craft

22%

30 real videos with sensible hand-coded features and age-normalised views.

Analytical honesty

22%

Correlations are real, outliers addressed, causation not overclaimed.

Actionable memo

11%

The 3 recommendations are specific to this channel and evidence-backed.

Channel ↔ CSV consistency

44%

The live channel and dated screenshot must match the CSV's videos, views and dates (allowing for view growth since collection). Invented videos/views or numbers that do not match the public channel are a fail.

How we grade your AI usage

30% of your score

Using AI is the point — it’s the skill this certificate proves. You’ll answer three short questions about how you used it: what you asked, what was wrong with its first answer, and what you changed. Specific, honest answers score high. “I pasted the brief and submitted the answer” scores near zero.