##Data Preparation: This data was scraped using python scripts from FesliyanStudios which provides you with royalty free musics in all trends that could be used in any of your projects or videos.

##Dataset Description: The original dataset contains 7 classes but this is just a subset of the original dataset which contains 2 classes. The audio files in the dataset were already preprocessed by converting them into smaller chunks of equal size. All music were divided into 10 seconds smaller chunks using FFMPEG.

The chunks of the same song have the equally spaced names like : Happy200_01-Happy200_16(Same song) The last two digits resembles the chunk number while the first three digits are the ID of the original song which is provided as the GroupID. The two classes are: 1.)Happy and 2.)Sad

MetaData.csv: File contains 6 columns with information useful for data analysis

GroupID : This column specifies the original song ID ImageID: This column contains the filenames(Chunks) Target: Specifies the classes to which a tune belongs to. Mean: Mean values of each audio file STD: Standard deviation of each audio file Skew: Skewness of each audio file