Interactive Reporting and Visualization Analytics

The main process of developing an analytical application involves iteratively deciding on the method used to perform a specific type of analysis, testing how well the method performs on test data and, if not satisfactory, either trying to optimize the method or trying different methods to build the application.

Interactive Visualization and Analysis tools can be used as an integral part of the InforSense platform, to enable an analytics professional to review the results of their analysis and to aid in the evaluation of the model.

InforSense provides Interactive Analytics Modules in four key areas:

• Data Exploration
• Classification and Predictive Modelling
• Clustering and Segmentation
• Text Mining

Data Exploration requires users to be able to visually explore data using different interactive charts. InforSense's Interactive Browser enables users to view the results of a workflow using: Scatter charts, Pie charts, Bar charts, Box plots, Line plots, Heat maps and Venn charts. Different views can be created of the dataset by combining different charts. However all charts are synchronized so that data selected in one plot is selected in all others.

Interactive analytics for Classification and Predictive Modeling are based around InforSense's Classification Studio. This studio provides a guided environment that simplifies and accelerates the development of high quality analytical workflows. Starting from "best practice" workflow templates, modeling wizards allow the analyst to easily customize settings and combine these with their own preferences. Alternative models can even be automatically created, enabling the user to explore a vast range of model settings within a few clicks. Users can compare the performance of these models side-by-side using diagnostic methods such as: Lift, ROC and Profit charts. The user is effectively provided with a model optimization environment. Workflows developed in a studio environment are encapsulated, allowing them to be easily embedded into other processes.

For Clustering and Segmentation-based analytics, InforSense provides the Clustering Studio. Using a similar approach to the Classification Studio, Clustering Studio provides a wizard to guide the user to customize best practice workflow templates to develop analytical workflows for segmentation analysis. The results of the segmentation are interactively displayed, graphically showing the make-up of the items in the cluster.

For interactively exploring textual data, InforSense provides the Document Browser. This provides users with a high level view of a document corpus, utilizing both ontologies and unsupervised hierarchical clustering of the documents. Users can drill down from this high level summary to read individual document texts, with important terms highlighted. There is also a free text search function and the ability to export selections made in this browser back into the workflow.