Reporting from International Information Conference on Search, Data Mining and Visualization (II-SDV) 2013 – Day 2
Copies of the presentations can be found here.
While I have been a long time attendee of scientific information meetings presented by Harry Collier and Christoph Haxel this was my first opportunity to attend the data mining and visualization conference. There were a number of things I found very appealing about the meeting. To start with, it’s focused on an area of high interest to me and therefore just about every presentation provided interesting insights. There were a little over a hundred people in attendance and all of them were involved, in some fashion, in the analysis and visualization community. This created a “cozy” atmosphere that was very conducive to carrying on conversations and provided detailed opportunities for networking whenever the plenary was not in session. I also very much appreciate that it is being held in Europe and is populated primarily by European practitioners. I find that I get a different perspective on this field from my colleagues in Europe and I always appreciate the unique insights I gain from being able to interact with them.
So, now that I have provided some color around the event itself, let’s jump into the brief summaries of the presentations from today:
Jean Archambeault, David Li Tang – National Research Council, Canada
(Panel Guests: Mike Alderton, Eric Fourboul – AMI Software, UK & France)
In Search of an Environmental Monitoring Tool: The Needs of a Research and Technology Organisation
Discussed a large, enterprise-wide system for current awareness and sentiment analysis for the National Research Council of Canada. The group wanted to continue with alerting services but was interested in providing more insight than data. They went through an extensive 72 point evaluation process to decide on a final product for implementation. In the end they selected AMI Software for the program. It scored well on the items that mattered most and has a multi-lingual interface.
Gerhard Fischer – Syngenta, Switzerland
Key Success Factors in the Setup of Cutting-Edge Patent Intelligence Services
Gerhard provided an excellent overview of the steps and skills required to establish a world-class technology mining process. The work conducted by the group is being utilized by new business units and decision makers within the company that were not previously clients. This was a model description of how patent information professionals are moving up the value chain in their company and providing insight to decision makers.
Dieter Küry – Novartis Pharma, Switzerland
Customized Newsletters – Strategies to Improved Current Awareness
This talk focused on the generation of current awareness services for a multi-national pharmaceutical company. They use a three-tier approach to allocated resources for supporting various user groups within the company. At the highest level, the reporting is customized specifically to the needs of the client and business. There was also a discussion on the development of a template for producing company-wide newsletters in various therapeutic areas.
Nils C. Newman, Alan L. Porter, Jon Garner – Search Technology / VantagePoint, USA
Tweet Mining: Is it Useful and Should we Bother?
Nils provided an overview of the difficulties encountered in getting access to a compressive backfile of tweets as well as finding a means of being able to separate the signal from the noise in this data. In the graphene area they were able to tease out some interesting ideas on thought leaders and trending topics. In nano-enabled drug delivery they were not nearly as successful mostly because of ambiguities in language and context. Nils’ assessment was that twitter mining might become useful in the future but it’s not quite there yet.
M Parthasarathy, J Cutie, Karthika Karindalam – SAARAA Medical Solutions, India
Pharmaceutical Companies and Social Media: Developing New Strategies
Discussed strategies for being successful in Social Media as a pharmaceutical company. They compared metrics around engagement and trust and how Pharma companies relate to one another in these areas. They also suggested that in order for pharma companies to be successful they should start early, start smart and don’t stop in order to achieve success in Social Media. It was also shared that customers tend to engage on therapies as opposed to brands or companies per se.
David Milward – Linguamatics, UK
Text Mining Diverse Data
Provided an overview of the methods developed from Linguamatics dealing with combing information from different sources as well as looking at second level connections. In this case a compound might be associated with a target in one reference while the target is associated with a gene or disease in a second reference. This approach allows association to be made between data from both references.
Diane Webb – BizInt Solutions, USA
Challenges in Visualizing Pharmaceutical Business Information
Diane provided two cases studies involving the challenges associated with developing new and useful visualizations associated with clinical trial data. Standardization of the data, consistency in the colors associated with the visualization, and thinking about the scale of the stages are critical design parameters. Eventually, the company would like to develop a “bulls-eye” visualization for clinical trials and have already developed a GANTT chart in conjunction with VantagePoint.
Martin Hofmann-Apitius – Fraunhofer Institute for Algorithms and Scientific Computing (SCAI), Germany
Text Mining at Work: Critical Assessment of the Completeness and Correctness of Knowledge-Based, Computable Disease Models for the Pharmaceutical Industry
Martin talked about text-mining in association with the Biological Expression Language for mining relationships within biological systems whether they be compound-target relationships, target-gene or gene-disease related details. He also mentioned that they have developed some scoring functions which allow them to test the statistical relevance of these relationships across a large number of documents.
Andrew Kirk – Visualizing Data, UK
Finding Stories and Telling Stories: Two Sides of Data Visualization
Andrew talked generally about the process of building impactful individual visualizations and infographics. He suggested that the user start with some thoughts on what the purpose of the visualization will be and, whether it is intended for the user to hear a story or find a story within the data. Another item to consider is whether the information is analytical or pragmatic in nature. Kirk also suggested that data presentation can be described with the five layers of a visualization – 1. spatial, 2. color, 3. interactivity, 4. annotation, 5. arrangement
Joachim Stellmach – European Patent Office, Germany
Graphical Representation of the Assessment of Inventive Step for Patent Applications
In this presentation a detailed discussion of the requirements for a European patent, novelty, inventive step and industrial application were discussed and the problem solution approach (PSA) was demonstrated as a means to adequately assess the inventive step for an application. The graphical representation was a means to lay out the components in the PSA process.
Zheng Li and Mark Atherton – School of Engineering and Design, Brunel University, UK
A Top-Down Method of Patent Mapping
Discussed the importance of looking at images in patent documents and how they can be analyzed to provide insight into the innovation process. There was also some discussion on a TRIZ approach to clustering patents but this area was not fully explored or described by the presenter.
I will end my coverage of II-SDV with the full collections of tweets that I gathered in association with the meeting in a subsequent post.