It seems like everything is producing data and we are surrounded by data from everywhere. Most of this data available on the internet is not in a structured form and not even in plain text. What to do with so much data? Are the data available in text or other form (video, images etc.) of any use? How do we utilize data that is available in different forms?
The advent of various social media sites and mobile apps boosted the production of data in various structures. Emails, Status updates, tweets, comments, blogs, and many other form social media exchanges create data that is unstructured.
One of the application researchers found that data in the form of social media exchanges was to identify “emotion” from rather than statistics. For example, if an author has written a blog on any topic then from its writing tone and style the views and emotions that writer want to showcase through its blog is identified. This emotion detection can also be applied on the comments of the readers on these blogs to identify their reaction to a particular blog.
If we just talk about the data that is available in the form of blogs; this data doesn’t follow any particular format or structure as each individual has its own writing style and structure. It was a big challenge to develop various algorithms that could work on such a diverse domain. But the researcher didn’t give up and many of them found amazing solutions and applications of such huge data sets.
Snapchat, a photo messaging application is very popular among the masses which enables its users to take photos, record videos, add text and drawings and share them with other users. In 2014 the average number of photos and videos shared per day on Snapchat was approximately 700 million. What can be done with so much data available in the forms of photos, images and videos?
It has been reported that Snapchat is currently developing a research team which will work on ‘deep learning’, branch of machine learning. The goal of these researchers will be to develop algorithms for deep learning analysis on images and eventually on videos as well. The main aim of Snapchat behind this new team development is to use deep learning concepts to improve the application or performance of its internal applications. Some other companies like Google, Microsoft, Twitter, Facebook, etc. are already in the process of building talent pools and Snapchat will be a good addition to this circle.