Description

• A business intelligence dashboard using Tableau
  
• A YouTube video metrics analyzer
  
  • Channel KPIs such as video count, likes, dislikes, views, duration, and video tags
    
• NLP sentiment analysis on YouTube comments
  

Introduction

This was inspired by a project I had previously worked on for a German-French media company as an analyst and consultant to help them make more data-informed decisions. This time, I wanted to take the data possibilities further with the power of Python and machine learning!

I used Python and Pandas to automate tasks to tidy data, unify it into one stakeholder reporting tool, and to use the data sources available to analyze content performance. I extracted channel and video statistics with the help of the YouTube Data API and performed sentiment analysis to find how viewers felt about videos and content.

Question

What makes a program a success?

A broadcast company can set itself apart by, instead of just broadcasting their own message, listening to what its viewers are saying. Machine learning can add value to improve content and uncover viewer sentiment. Natural language processing on video comments reveals positive, negative, or neutral attributes.

Installation and Requirements

You need the YouTube API key for this (I don't include mine)
https://developers.google.com/youtube/v3/getting-started
https://console.developers.google.com/

Viewer Metrics Dashboard

Sentiment Analysis on YouTube Comments

Bigram to count the frequencies of 2 consecutive words in comments

Weighted words in video comments

Dashboard with multi-channel metrics

Tech Stack

Python, pandas, numpy, sklearn, nltk, vadersentiment,TabPy, Tableau