Format

====================
1 Posts DataSet
====================

This is in CSV File Format. Contains basic information about post, language and post types as listed here:

permalink: Sharechat post index ID.
createdAt: date of post.
itemType: type of post (text, image, video etc.).
likes: likes on post when fetched.
views:	views on post when fetched.
shares: shares to Whatsapp when fetched.
tagText: similiar to hashtag, used while making a post. Eg. #GoodMorning
tagHandle: ID of tagtext.
language: language of user, selected from one of the 14 language forums while registering.


====================
2. Images Dataset
====================

Image hashcodes and permalink (ID to refer post information as in above (1))
------
This is in CSV file format. Contains the Facebook PDQ image clustering headers for images. For each image post the dataset contains the following:


id: cluster ID as assigned from PDQ (similar images will have same cluster ID) to image. Please check paper for method we choose for distance and clustering.
size:  cluster size of cluster ID of which image has been grouped. 
hash: Hash codes generated from PDQ.
norm, delta, quality, dims, readSec, hashSec: features set and obtained during computation of hash codes.
permalink: index ID to link image data to posts data above (1). 
ocrText: Optical character recognition from Image text.