on signup

create table user (details... ,uid, categories)

on each post by any user

create table posts(pid, category, content, title, *media(url)*, likes

create a category table for each category (for easy routing of the “category-sise-posts section

also create a user-category table with primary key category

create category-table(category, no_of_likes_by_user, no_of_posts_in_category,weight)

Flow of events :

Extraction of posts
Creating pseudo Table
Assigning weights
Sort

1. Extraction of posts

on the basis of timestamp, extract posts which are one month/week (to be decided) old and assign weights to them. 
Display rest in a reverse chronological order

1. Creating a pseudo Table

From the extracted posts, create a pseudo table

1. Algorithm for assigning weight to extracted posts

3.1 On creation of a user account, 
    n=no_of_categories_selected
    category-table.weight = (100/n)

3.2 after some user interaction with the system
    change_Weight_Each_Day(category-table.no_of_likes , category-table.no_of_posts)
    {
      for each category:    
        category-table.weight-new = 0.9*category-table.no_of_posts + 0.1*category-table.no_of_likes                                                                                                                    
        category-table.weight = 0.8*category-table.weight + 0.2*category-table.weight-new
     }

1. Sort the extracted posts of the pseudo table (pid, final_weight)

   {
       for each pid
       {
           st = pid.category  						                      //from posts table	
           category_weight = category-table.weight[st]
           final_weight = pid.likes * category_weight                 
       }
   
       select * from pseudo table order by final_weight.