Analysis of Multi-Modality Contents in Data for Visualization of Events

Any Social media documents or any document collected from internet contains multiple stories attached with it. Especially having images concerned with that story. Searching sorting and analyzing this huge amount of documents is quite tedious job for user. Because it is very time consuming to sort the documents containing particular story and clustering of such multiple events is time consuming job for analyzers. Modeling these events according to their year and type is useful for easily sort the information among the huge data. So the detailed analysis of multi-modality contents in the dataset, the methodologies for modeling visual representative topics and text oriented topics together is also presented in this paper. Multi-modality contents can model using the topic allocation model but studying semantic relationship between them is again the difficult job. So semantic relationship is also highlighted in the paper so as to model the information and images efficiently. One part is detect and scrutinize dense block of data which is worth inspecting, typically indicating fraud and sort them in the order of importance (suspiciousness) is also described. In multimodal visualization with help of Graphical representation such as pie charts, Bar graphs the comparison between the various events can be easily done. The workflow is presented and finally conclusion is presented along with future work.