We live in the age of social media. Approximately 6,000 tweets are tweeted every second!
It’s no wonder that social media has begun to affect more than how many cat videos you watch a day. In the same way that the debut of the home television affected the 1960 U.S. Presidential election, social media has become an entirely new variable in the complex equation of public decision making.
As exciting as it is to be living in such a connected time, it’s even more exciting to have access to tools that let us, everyday people, explore and catalog and observe the 6,000 public view points being expressed every second!
Using a combination of Twitter’s API, IBM Watson’s Sentiment Analysis, and Initial State’s data visualization all working together through the PubNub Data Stream Network, I decided to take a look at how people feel about the current U.S. President, Donald Trump. This provides a way to look at popular opinion on your own and also seems to validate the idea of negativity bias.
I’m going to walk you through what I’ve done so you can perform your own experiments!
In this tutorial you will:
- Use PubNub’s stream consuming the Twitter API to filter tweets by keywords
- Use PubNub blocks to send tweets through IBM’s Sentiment Analysis
- Send Sentiment Analysis results through blocks to Initial State’s real-time dashboards
- Build an informative and interactive dashboard
Note: This is not meant to be political in any way! It is meant to be a fun and informative project looking into public opinion on one of the nation’s current hottest topics. Also, apologies in advance for any crude tweets you might see – the code could definitely be edited to filter out tweets with certain content.
This tutorial is part of the series, “Learning How to Build Real IoT Applications” >>