The human mind has a very weird way of associating images with words. For instance, if someone told you the word heaven you are likely to see heaven instantly, based of course on the popular imaginary representations of heaven. The same goes with data science. You may never have been to a data science lab but you have an image. It could be the picture of giant plasma screens and bespectacled men and women typing away. These images influence our opinion of a certain thing, and more often than not we miss the fun side of it. This article focuses on the fun side of data science.
The obvious grounds for data science
Learning data science is not the prerogative of scientists and analysts any more. A financial advisor, a sports manager, a media consultant, a linguistic expert, everyone can have some use of a data science course. You can read more about it here. It is pretty obvious that a finance and banking corporation would use historical data to find out how likely a person is to be a loan defaulter, or that an Ecommerce firm would want to deploy analytical models to understand consumer behaviour. Our quest is different here.
Data science in music
You knew I would get into this, didn’t you? Music as it has been so far is an infinite field in terms of emotions expressed but objectively it is down to twelve notes, few time signatures, and certain proven formulas for composition. It can be narrowed down to several thousand possibilities. And that is not too much for a machine learning algorithm. There have been successful efforts to create AI driven composition of notes and patterns that sound like real music.
Writing a novel
First you expose a system of deep learning networks to a few hundred books that have been written by human beings. Then you expose it to an audio visual experience. Say you take it to a drive and let it absorb the environment. Obviously there has to be some supervised learning to help it recognize the things. It can actually mash up the literary and linguistic knowledge with the audio visual experience and create original content. Right up from the scratch. It has been done and will be done with more success.
Creating new cooking recipes
There could very well be a data science cuisine in near future where machine learning systems create the recipes. There have been successful experimentation in this line. You feed the program with recipes that include the flavours and aromas created by each ingredient. So, you perform some unsupervised learning. And the program will belt out original recipes. Some of them have actually impressed the scientists a lot.
Casting for movies
Casting for a movie that has a considerably large cast is a nightmare. Assuming the suitability of a person for a role while taking into account the profitability of that choice in the box office and then being completely wrong about it. Data science has put a happy end to this predicament. The largest production houses around the world are using data science for casting. The analytical models analyze the scripts and draw references from the data pool regarding actors, their previous works and box office success and provide the best possibilities.