In this project, we collected data from YouTube channels that were named in Top Programming Guru. These YouTube channels helped people to advance their careers in programming. The ranking of this YouTube channels is based on the vote of the community.
The steps of the project:
- Get the Top Programming Guru channels
- How to get a YouTube channel Id?
- Extract YouTube channel statistics
- Analyzing the data set of YouTube channels
- Extract YouTube video statistics
- Analyzing the data set of YouTube videos
- Analyzing the data set of YouTube playlists(to do)
- Create database from csv file(to do)
Python Version: 3.7
Packages: pandas, numpy, sklearn, matplotlib, seaborn, selenium
Python YouTube API Tutorial: Getting Started - Creating an API Key and Querying the API
Using Python and YouTube API to Create Analytics on any Channel.
We scraped the page Top Programming Guru to get the nominated channels. For each channel, we got the following:
- channel name
- url
- rank
We used the YouTube API to collect the following data for each Top Programming Guru channel:
- channel name
- channel title
- channel Id
- description
- rank
- country
- view count
- subscriber Count
- video count
- published At
- uploads
Then, we used the YouTube api to extract statistical
data for each video and playlist in each channel.
For each video, we got the following:
- video Id
- title
- tags
- view count
- like count
- dislike count
- comment count
- description (to do)
- duration
- channel id
- default Audio Language
- published At
For playlist, we got the following:
- playlist Id
- title
- description
- itm count
- publishe at
- channel id
Finally, we retrieved for each playlist its items using the playlist Id.
- video Id
- playlist Id
After collecting the data, We needed to clean it up so that it was usable for our model and visualization . we made the following changes and created the following variables:
- Removed duplicates channels
- Removed outliers
- Made a new columns for location
- Made a new columns to measure interaction with videos
- Added a gender column for each YouTube channel
- Used data from the first video uploaded to a channel to estimate its age. (to do)
We looked at the distributions of the data, and the value counts for the various categorical variables. Below are a few highlights from the pivot tables.
