Skip to content

HN/Reddit social news websites big data analytic with Hadoop Hive and MapReduce. Discover correlation between a website's content and number of votes it receives on social news websites such as Reddit and HackerNews.

Notifications You must be signed in to change notification settings

mutaphore/social-news-bigdata

Folders and files

NameName
Last commit message
Last commit date

Latest commit

 

History

77 Commits
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 

Repository files navigation

CONTENTS
--------

/crawler
Python scrapy web crawler responsible for crawling urls retrieved from HN/Reddit
and save the content of those webpages as well as various metadata.

/datumbox
Machine learning framework from datumbox.com. The framework can be used
to perform many types of analysis such as sentiment analysis, topic
classication, keyword extration etc. on a large dataset.
Currently our project is not using this framework; however it can be an
additional add-on analysis toolbox later if we want to do a more
in-depth study.

/hackernews
hackernews_api.py - this script makes requests to the HN api and downloads all the data from their RESTful service.
extract_url_from_csv.py - extracts urls from the csv file generated by the HN api. This is used for testing only.
get_urls.py - extracts urls but check if they are valid by making a request. Also for testing.
urls.txt - example of some urls retrieved


/hive
process_hn_api_output.hql - cleans the HN api output file to remove commas, NULL fields, etc. for easier processing later
process_hn_crawl_output.hql - cleans the HN crawl output file to remove commas, NULL fields, etc. for easier processing later
filter_hn_api_fields.py - this script contains the transform function used by process_hn_api_output.hql 
filter_hn_crawl_fields.py - this script contains the transform function used by process_hn_crawl_output.hql
join_output_and_crawl.hql - joins the HN api dataset with crawled dataset so we can get further correlation info
get_reddit_output.hql - cleans and export reddit dataset to local directory
remove_fields_comma.py - an udf used by hive to remove comma from title field so that data can be process by mapreduce job 

/mapreduce
hn_num_comments - MapReduce program to get relationship between number of comments received on a post vs. the average score received
hn_page_content - MapReduce program to get relationship between number of links, scripts, images, styles vs. the average score received
reddit_post_type - MapRedue program to get relationship between the subreddit of a post on Reddit vs. the average score received
reddit_post_hour - MapReduce program to get relationship between the hour of a day vs. the average votes received

/pig
filter.pig - filters the HN output files to only get the "story" type urls.


/reddit
cleaned_data - the final reddit dataset ready to be consumed by the reddit_post_hour program
getposts.py - depricated
output.csv - raw reddit post data collected by calling reddit api from reddit.py script
post_datastructure.txt - descriptor for the reddit object return by the reddit api
reddit.py - python script that call the reddit api to get the raw reddit dataset (output.csv)

About

HN/Reddit social news websites big data analytic with Hadoop Hive and MapReduce. Discover correlation between a website's content and number of votes it receives on social news websites such as Reddit and HackerNews.

Topics

Resources

Stars

Watchers

Forks

Releases

No releases published

Packages

No packages published