Using Pandas to optimize YouTube Channel Growth

What makes a successful coding channel on YouTube? Let’s analyze the data from some of the most successful programming channels to find out. We examine metrics like views per like and engagement patterns to identify what actually works.

Data Analysis Process

Initial Data Preparation:
– Collected channel data including video titles, views, likes, and comments
– Cleaned the data by removing irrelevant columns
– Converted metrics (like “1.5K” to 1500) to numerical values
– Filtered out videos with zero likes or comments

Key Performance Metrics

Two main metrics were analyzed:
– Views per like: How many views it takes to get one like
– Views per comment: How many views it takes to get one comment

These metrics help identify truly engaging content versus videos that might get views but don’t inspire interaction.

Key Findings

1. Channel Comparisons:
– Free Code Camp: 400,000 average views
– Programming with Mosh: 900,000 average views
– NeuralNine: 36,000 average views

2. Content Insights:
– Beginner content dominates views (top 3 videos: 44M, 39M, 17M views)
– Long-format videos (1+ hours) perform surprisingly well
– Intermediate content gets significantly fewer views (around 3M)

3. Engagement Patterns:
– Smaller channels often have better engagement ratios
– Views don’t always correlate with likes and comments
– Some videos with fewer views have higher engagement rates

Surprising Discoveries

Long vs Short Format:
Top-performing videos are mostly long-format (1-15 hours). Short-form videos (under 60 seconds) didn’t make the top rankings when filtered by views.

This is unexpected given the conventional wisdom about decreasing attention spans. Apparently, people are willing to invest serious time in comprehensive tutorials that teach them something valuable.

The Beginner Content Gap

The data shows a massive drop-off between beginner and intermediate content. Beginner tutorials get 10x or more views compared to intermediate topics.

This makes sense when you think about it – everyone starts as a beginner, but not everyone continues to intermediate levels. The audience naturally shrinks as content gets more advanced.

Practical Takeaways

For Content Creators:
– Beginner content has the largest potential audience
– Long-format comprehensive tutorials can perform well
– Focus on engagement, not just views
– Consider the natural drop-off between beginner and intermediate content
– Smaller channels can compete through better engagement

For Learners:
– Comprehensive long-form tutorials are worth your time
– Popular doesn’t always mean better – check engagement ratios
– Smaller channels often provide more focused, engaged communities

Download the Analysis

Want to perform your own analysis or see the detailed code? Download the complete Python scripts and data here:

You can adapt this analysis to study your own channel or compare different niches.

---

Want to improve your Python data analysis skills? Check out our courses at Training Scientists for in-depth tutorials on scientific computing and data analysis.