Ecosyste.ms: Awesome
An open API service indexing awesome lists of open source software.
https://github.com/markledwich2/TransparencyTube
https://github.com/markledwich2/TransparencyTube
Last synced: 3 months ago
JSON representation
- Host: GitHub
- URL: https://github.com/markledwich2/TransparencyTube
- Owner: markledwich2
- License: mit
- Created: 2020-09-09T23:34:44.000Z (about 4 years ago)
- Default Branch: master
- Last Pushed: 2023-02-18T01:35:05.000Z (over 1 year ago)
- Last Synced: 2024-04-18T00:22:20.044Z (7 months ago)
- Language: TypeScript
- Size: 18.9 MB
- Stars: 9
- Watchers: 4
- Forks: 1
- Open Issues: 0
-
Metadata Files:
- Readme: README.md
- License: LICENSE
Awesome Lists containing this project
README
transparency.tube[![Netlify Status](https://api.netlify.com/api/v1/badges/5c63003a-7edb-41e3-ac4d-221253a66cae/deploy-status)](https://app.netlify.com/sites/laughing-hopper-65d275/deploys)
## Download Data
Data is collected by [recfluence](https://github.com/markledwich2/Recfluence).Download latest classified [channels](https://ttube.azureedge.net/public/results/ttube_channels.jsonl.gz)
For video statistics, captions and other larger please contact [email protected] to arrange at-cost access to the database.
## Metrics
**Video views**: *Transparency.tube* records statistics for younger-than-365-days videos for all channels each day.Note:
- For channels with large back-catalogs, we read older videos to try and capture all the views on a day
- When new channels are added to transparency.tube, the history of views are estimated based on the upload date of videos.Click on a channel to see more detail about the collection of video statistics.
**Channel Views**: The total number of channel views as reported by the YouTube API.
Note:
- More reliable than `video views` for channels with large backlogs (e.g. AP Archive), but less reliable for channels that have videos removed.
- We only have data for channels since we have added them to our dataset, so we show only the latest figures.**Watch Hours**:
This estimate of hours of video watched is based on the [data collected](https://github.com/sTechLab/YouTubeDurationData) from [this study from 2017](https://arxiv.org/pdf/1603.08308.pdf). For each video, we calculate \`*average % of video watch for this videos duration\` x \`video views\`**Subscribers**: The number of subscribers to the channel as provided by the YouTube API. Note, a small amount of channels hide this data.
## Data Collection and Periods
Stats are collected each day starting at 00:00 GMT and the dates displayed are all in GMT. We decided to use a common date because this makes it easy to compare full-days that are the same for different timezones.## Classification
For a description of how channels are manually reviewed, see [Recfluence](https://github.com/markledwich2/).
To understand the automatic classification, see [Chan2Vec](https://github.com/sam-clark/chan2vec#chan2vec)