Data Exploration and Analysis

a web-based system of interlocking tools for quantitative analysis


The interface is a browser-based, thin client, with the data remaining in an online repository, and the statistical modeling occurring on a remote server.

A core concept of TwoRavens is mode. Different modes signify different main goals a user might be attempting in the current part of their analysis. When the system is in different modes, the features available to the user are swapped or modified to better work for that intended task.

Dataset Mode gives different ways for a user to bring data into the system. Here a user can upload new datasets, switch between datasets already available, and access the Datamarts. Datamarts are a way to augment current datasets with additional relevant data to a problem.

Explore Mode allows a user to perform exploratory data analysis by intuitively generating visualizations of relationships between variables.

Model Mode allows a user to construct models of interest, specify a target to predict, explore the automatically discovered problems suggested by the system, override any metadata created for features and change the problem characteristics of the model.

Results Mode allows a user to automatically generate machine learning solutions for a user constructed model, explore previously constructed models to gain insights into model performance, and understand the relative importance of features within and across different machine learning solutions.