==============================
This is project created by Siddharth Boobna, Yash Parikh and Prateek Sinha for Big Data Analytics.
Department of Computer Science
Columbia University
ssb2171 [at] columbia [dot] edu
Department of Computer Science
Columbia University
yp2348 [at] columbia [dot] edu
Department of Computer Science
Columbia University
ps2791 [at] columbia [dot] edu
-
After downloading the source code from here and the dataset, you must first run the data extractor files found here
-
The data extractor will create CSV files which can be used with Hive. Note: You may have to modify the extraction scripts to give the path and filenames of your choice. Also, you can modify the columns being selected to work with any dataset of your choice.
-
After the data has been extracted, you may run the Hive Script found here which contains the recommendation algorithm.
-
The output of Hive will go to S3. You will have to modify it to target your S3 bucket.
-
Next, you need to modify the REST_API here
a. You need to modify the RecoLister.java file to add your own AWS credentials and also the change the path in the bucket where the output of Hive goes.
b. You can then upload the war file to your EC2 and hit the url http://your-ec2-url/RequestHandlerServlet?uid=some-uid-here&numBusinesses=5 to get the appropriate number of recommendations for that user.
c. The result will be ontained in a json format and can be used with any UI to be displayed in a meaningful manner.
-
You can also use our UI code which maps our recommendation and displays a ranked list on side.