Summary:
- The blog emphasizes the limitations of existing research methodologies, particularly in social media platforms like Reddit and Twitter, where access to denominators has been restricted.
- It introduces the InnerTube API, an undocumented API for YouTube, and describes the "drunk dialing" approach, explaining how random YouTube video URLs were generated to estimate the platform's size.
- The author acknowledges the method's imperfections but highlights its significance in providing a unique, large-scale, and relatively unbiased dataset for studying YouTube content, including aspects like language diversity, video age, and viewership distribution.
- The post mentions ongoing research plans, ethical considerations regarding sharing low-popularity public videos, and the intention to maintain Tubestats as a valuable resource for understanding user-generated media platforms.
- Reader comments reflect various perspectives, including discussions on the methodology, typos, and implications of the research findings, adding depth to the conversation.
- The author discloses a personal note about having COVID, which might impact the coherence of the post, inviting engagement and understanding from the audience.
Discussion (1)
Interesting how much growth in content in just this year alone