Aravind Mohan

Research Interests:
  • developing scalable and usable data management tools and data models leveraging the cloud computing environment to address the five “V” challenges of big data: volume, variety, velocity, veracity, and value.
  • developing high performance scheduling algorithms for big data workflow optimization driven by Quality of Service (QoS) in a heterogeneous cloud computing environment.
Research Projects:

 DATAVIEW (http://dataview.org/): An online big data workflow management system (BWFMSs), that enables domain scientists to create, edit, search, share, debug and run big data workflows online. A big data workflow represents a multiple-step data analysis pipeline by chaining several data analysis modules together via data links that is used to connect the output of one analysis module to the input of another analysis module. It focuses on large-scale data processing and analytics with a “scale-out” architecture and a “moving-computation-to-data” processing paradigm. (Download DATAVIEW)

 *We use FutureSystems and AmazonEc2 as our cloud platform.

              A Sample Big Data Workflow in DATAVIEW.

 

 Opinion Mining : Our Big data workflow technique is applied to:
1) first, develop crawl workflows to collect the large scale web data with user generated opinion text along with their meta data (such as gender, age and location) in several domains.
2) second, develop opinion mining workflows(by using topic modeling techniques) to detect aspect-level sentiment analysis using the user meta data.

 *Our dataset (sample here) is well appreciated by the IR community. If you are interested in using the dataset for your research, please feel free to contact.