I want to grab youtube comments, good to feed into mindmanager?
hi,
sometimes, the comments are more gold than the video itself!
chatgpt allowed me a lot of things i could not do before.
to save time from "get lost"in youtube,
i think someone can:
disable youtube on every PC, only allow one connected but without browser.
install some downloader, ytd-dl is foss, but i currently bought byclick downloader.
indeed at this step, you can use ffmpeg to extract sound file.
put into whisper API for speech to text STT
obtained SRT for subtitles, or text (i prefer keep both).
then feed into chatgpt to summarize the txt.
but other then the content,
there are websites that let you read the "metadata", the view count, date, channel subscriber etc.
then the above comment TREE. TREE form is best using mindmanager, i think better than excel.
i am just thinking in the TREE of comments, how to also state the commentor's name, time, "
ägree" times etc. reply from the OP likely will need specially highlighted too.
may be could also ask gpt to summarize ALL comments. and esp highlight those highlighted by human.
in this will i think it's easier to go thru the comments.
btw, chrome already have comment analyzer but seems less than the above.
text form:
The individual, Krsto E., is interested in extracting valuable information from YouTube comments, which they find often more insightful than the video content itself. They propose a multi-step process to efficiently collect and analyze these comments without getting distracted by browsing YouTube. The process involves disabling YouTube access on all but one computer, which would not have a browser, to prevent aimless browsing. They suggest using a downloader like ytd-dl (a free, open-source tool) or byClick Downloader to download videos. Then, leveraging ffmpeg to extract the audio from these videos and using Whisper API for speech-to-text conversion to obtain subtitles or plain text. This text can then be summarized using ChatGPT to extract the essence of the video content.
Additionally, they mention utilizing websites that provide video metadata (such as view counts, upload dates, subscriber counts, etc.) and suggest that organizing comments in a tree structure (using MindManager or a similar tool) could provide a clearer overview of the discussion. They believe this method, especially if it includes details like the commenter's name, time of comment, number of likes, and special highlighting for original poster responses, would enhance understanding and analysis of the comments.
Krsto E. also suggests asking GPT to summarize all comments, particularly emphasizing those that have been highlighted by humans, to make it easier to navigate through the comments. They note that while Chrome offers a comment analyzer, it appears to be less comprehensive than the proposed method.
mindmap form:
- Krsto E.'s YouTube Comment Analysis Proposal
- Goal: Extract valuable insights from YouTube comments
- Process Overview
- Disable YouTube access on all PCs except one
- The exception should not have a browser to prevent distractions
- Use a downloader to obtain videos
- Suggested tools: ytd-dl (FOSS) or byClick Downloader
- Extract audio from videos
- Tool suggested: ffmpeg
- Convert speech to text for subtitles/text
- Suggested API: Whisper
- Summarize text content using ChatGPT
- Use websites to read video metadata
- Information includes view count, upload date, subscriber count, etc.
- Comments Organization
- Suggests using MindManager for a tree structure
- Include details in comments tree
- Commenter's name, time of comment, like count
- Special highlight for original poster responses
- Proposes GPT summarization of all comments
- Highlight comments emphasized by humans
- Tools and Enhancements
- Chrome's comment analyzer mentioned
- Considered less comprehensive than proposed method
text form:
The individual, Krsto E., is interested in extracting valuable information from YouTube comments, which they find often more insightful than the video content itself. They propose a multi-step process to efficiently collect and analyze these comments without getting distracted by browsing YouTube. The process involves disabling YouTube access on all but one computer, which would not have a browser, to prevent aimless browsing. They suggest using a downloader like ytd-dl (a free, open-source tool) or byClick Downloader to download videos. Then, leveraging ffmpeg to extract the audio from these videos and using Whisper API for speech-to-text conversion to obtain subtitles or plain text. This text can then be summarized using ChatGPT to extract the essence of the video content.
Additionally, they mention utilizing websites that provide video metadata (such as view counts, upload dates, subscriber counts, etc.) and suggest that organizing comments in a tree structure (using MindManager or a similar tool) could provide a clearer overview of the discussion. They believe this method, especially if it includes details like the commenter's name, time of comment, number of likes, and special highlighting for original poster responses, would enhance understanding and analysis of the comments.
Krsto E. also suggests asking GPT to summarize all comments, particularly emphasizing those that have been highlighted by humans, to make it easier to navigate through the comments. They note that while Chrome offers a comment analyzer, it appears to be less comprehensive than the proposed method.
mindmap form:
- Krsto E.'s YouTube Comment Analysis Proposal
- Goal: Extract valuable insights from YouTube comments
- Process Overview
- Disable YouTube access on all PCs except one
- The exception should not have a browser to prevent distractions
- Use a downloader to obtain videos
- Suggested tools: ytd-dl (FOSS) or byClick Downloader
- Extract audio from videos
- Tool suggested: ffmpeg
- Convert speech to text for subtitles/text
- Suggested API: Whisper
- Summarize text content using ChatGPT
- Use websites to read video metadata
- Information includes view count, upload date, subscriber count, etc.
- Comments Organization
- Suggests using MindManager for a tree structure
- Include details in comments tree
- Commenter's name, time of comment, like count
- Special highlight for original poster responses
- Proposes GPT summarization of all comments
- Highlight comments emphasized by humans
- Tools and Enhancements
- Chrome's comment analyzer mentioned
- Considered less comprehensive than proposed method
ps, the comments need to be retrieved with youtube DATA api thru python.
it's the most difficult part, the metadata like dates are more easy and already done by some websites.
ps, the comments need to be retrieved with youtube DATA api thru python.
it's the most difficult part, the metadata like dates are more easy and already done by some websites.
---