Entries.Ortiz2010 History
Hide minor edits - Show changes to markup
V´ıctor Pascual Cid | Bestiario |
Víctor Pascual Cid | Bestiario |
\\
\
Use cases of Impure, and information interface
Santiago Ortiz | Bestiario |
V´ıctor Pascual Cid | Bestiario |
Published: October, 2010
Full article |
Summary of the Tool
Impure is a visual programming system that enables users to gather, manipulate and visualize many types of data. Using a node link approach where nodes are data operators or visualizations and links represent data flow, users can construct simple and complex mashups to analyze and understand data from many different sources. The main advantage of Impure is the possibility to use the visualizations as dynamic filters of the data, that can feed other modules, converting visualizations in part of the whole information flow rather than only acting as interpretation tools.
Summary of its Impact
From the collaboration with two major companies, we describe below two examples of impact.
Study of brand perception
The Havas Media lab wanted to understand how customers and potential customers refer to different products and brands and how this relates to sales behavior. Without the intervention of Opinion Mining agencies, their analysts were able to generate an Impure workspace that allows to visualise and explore the number of times that relevant sentences appear in the results from Google regarding two main brands. The interactive space provided by Impure has also allowed the analysts to generate the senteces to be searched interactively, by using Impure modules to generate and visualise semantic networks of adjectives that will be finally part of the sentences to search. The final results can be explored with another visualization that compares the proportion of occurrence for each phrase for each brand.
“Before the existence of Impure, we had to contract several Opinion Mining agencies to extract online brand reputation. Impure has provided us an easy way to access, analyze and visualize such data without the need of such specialized agencies, reducing costs and increasing response time. Furthermore, Opinion Mining agencies are usually very topic specific, generating problems when comparing data from different types of markets.”
Study of correlations values of listed companies
Repsol's Department of Operational Models wanted to compare their stock market behavior with that of other companies. They developed an Impure workspace that collects such information and calcaulates correlations among the data, which is visualisaed with graph-based layouts. A stacked graph also allows to understand stock values with time.
“The fact that Impure enables to represent and directly manipulate the data in a visual way has allowed us to communicate our analysis results in a better way to the non-expert board of directors, facilitating the dialogue with them. Moreover, Impure helps in the process of building very specific dashboards to analyze the data. Such dashboards were previously ordered to third party companies that will be no longer needed. From our experience, classic analysis tools present input and output processes, where the data transformation is obscured to the end users. Impure provides ways to create, follow and control the whole data flow, enlightening the analysis process.”
Links
- www.bestiario.org
- www.Impure.Com
- Demo video
Copyright of article and images belongs to the author or his employer unless otherwise mentioned