| Web3.0: Social, Services, Sensor Semantic Web - Working on a part of Citizen Sensor project where human being’s act as sensors. The main objective of this project is to understand the events from the perspective of general citizens. Basically involves extracting tweets from twitter for a specific domain, Financial Crises in our case, looking for interesting concepts in the tweets, searching for those concepts in a rich domain model like an Ontology (carved from Wikipedia in our case) to get more information about the concept. Finally semantically annotating the concepts using the Ontology classes which would help understanding the picture from an individual’s perspective and lead to applications like trend analysis etc. My work in the project deals with providing interfaces between the tweets extraction and the RDF store containing all the instances related to a specific domain. The instances were taken as a dump from Dbpedia. The interfaces were built using the Sesame RDF API. Research oriented challenges involved taking 2 terms retrieved from a tweet and looking for relations present between the terms in the RDF store which could be multiple hops away. | Winter 2009 |
| Semantic Web - The major goal of this project was converting Observation and Measurements Language an OGC specification to Resource Description framework. Observation and Measurement Language (XML language) encodes data within XML nodes having the same name. The problem with approaches looked at then converting XML to RDF, was that they did not allow using XPATH predicates to map to the exact XML node. The tools allowed mapping to the topmost class in the hierarchy and generated new Ontology Schema based on the XML schema provided. Hence a JAVA approach was taken where the data required from O&M was parsed out and looking at the structure of the existing Ontology Schema and a Mapping file an RDF Instance was created. More information about this can be found on the Research Page. | Fall 2008 |
| Web Information Systems - Whooflix, Facebook project where the idea was to model a user in social space using multi dimensional approach. Aided by shallow semantic models that describe each dimension, a network can be expressed as a space of users with links existing between users at different dimensions. The strength of the relationship can also be modeled along each dimension. In Whooflix we looked at the dimension of movies where my work was to extract movie related data from IMDB like movie names, actors, actresses, genre, directors and store the data in a MySQL database. The first step was to understand the URL patterns for the movie names and search for a root page to start the extraction from. Then using the page containing the year information as the root page, branching out to get the information related to movies. The entire framework for extraction was built in JAVA. | Spring 2008 |