Analyzing Social Media
We used a neural network to look at 1,000 images of an Instagram timeline and used the data to create a word cloud of the most popular hashtags. To be sure, the program had no access to meta data, tags, or descriptions of the photos, and was only able to analyze the images themselves.
For this we used cognitive services from one of the cloud computing providers to get tags and descriptions of each image. Before we get into the details, let us have a look at the word cloud created by our program.
To accomplish this, we sent each image to the provider's cognitive services via an API call and received a response that included tags and descriptions of the image. We then took the most popular tags and created a word cloud. The service would also come back with a confidence score for each tag, which we used to eliminate photos the AI was not very confident about. For example, it would classify a lake of turtles as "a bunch of bananas" with a low confidence score.
This is just one example for using AI for social media analysis. Other use cases for this could be finding the most popular products in an Instagram feed, analyzing the sentiment of tweets, or finding the most popular locations in a set of photos. Trending topics on social media can also be analyzed with AI to see what people are talking about. You could also use AI to generate descriptions for products or services, or to find the best time to post on social media. Responsible use of this technology can help you better understand your customers and what they want. The metaverse is the next level of social media where you can use AI to create a virtual world that is responsive to the people who inhabit it.