MyScienceWork @ ConTech London 2018

Editor's Notes

• #3: MSW aims to make science more open and accessible to all. This is reflected through 3 pillars: Popularization and scientific communication: we work with a network of scientific journalists to popularize and enhance scholarly outputs + Access to scientific publications through our open database of more than 70M publications and patents Polaris OS: which aims to provide research institutes with an expert solution in data management. Sirius: Polaris OS goes hand in hand with the data processing expertise we have developed over the past 8 years.
• #5: Let’s talk about how to build an open repository for scientific institutions to analyze and process data
• #6: Polaris OS is an open source solution developed in partnership with the famous French Institute for Demographic Studies INED. We started the project with them last year and at this time, they would like to get an institutional repository solution that can address this specific challenges: Provide their researchers with an easy solution for publication deposit: UX/UI, auto-completion of metadata and useful tools (bibliographic management, push of publication toward other repositories…) Get a solution where they can control the quality of data produced by their research activities and thus be able to: Increase the visibility of their research outputs in other databases (national or thematic repo…) and on the web Produce analytics graphs and reports for evaluation on solid basis Get a sustainable solution: Metadata model flexible... Based on open and innovative technologies (data oriented…) Interoperable with other framework and infrastructures Easy new development environment Being able to be as autonomous as possible to manage the solution (no IT expertise required).
• #7: When we started to think about the project, we identified the main challenge as having a solution data oriented that can guarantee a sustainable evolution.
• #8: In order to help institutions overcome technical and budgetary hurdles, we developed an open source repository designed to analyze and process data. This is a major technological breakthrough to improve data management, analyze research impact, and better user experience. Polaris OS integrates into the ecosystem of research institutes by recovering/exporting all types of data (scientific, technological, financial, and managerial) to and from their existing information systems to organize, refine, and enrich it. Polaris OS is a combination of a: Enterprise Resource Planning: Integrate external streams: it’s the ability of interfacing the solution with external sources coming from IT department from our clients. Aggregate data under a common model: it’s the ability to gather data disseminated on various databases. Roles & users management: defining fine-grained access to the platform on both backoffice & frontoffice Extract Transform Load: Using pipeline to handle stream of data. A pipeline is a succession of functions used to format, complete, filter, transform as well as validate an input. The pipelines are extremely flexible and give access to a wide range of transformations and completions. Complex functions can be designed using transducers (a combination of simple functions that output always the same result given the same input). Data can be retrieved from a wide range of protocols including : ODBC (connection to relational databases), SOAP-XML (used primarily with Java and Hibernate), REST (standard way to exchange data on the internet nowadays), SFTP ((secured) file transfer protocol). Content Management System: Templates describe the way a page looks (does it have a header, a footer, a horizontal or vertical menu, …) Menus can be changed (items can be added, moved, removed, …) A widget is a basic element that has a unique purpose: search, browse, showing an image, a text, and so on. Widgets are placed on a 12-column grid based on standard CSS framework used by web designers and integrators.
• #9: I would like to emphasize on some few matters that make Polaris OS unique. First of all, we choose very carefully the technology necessary to build the solution with 2 ideas in mind: the technologies chosen must be open and innovative (data oriented). As an example, we choose: Vue.js: this is one of the most promising technology to develop reactive interface Elasticsearch: currently the best search engine for data Node.js: framework…
• #10: Second point, Interfaces have to be designed for all type of users. In our case, we identified 4 types of users and thus we set up 4 customized interfaces.
• #11: Get a solution that allow the institution to get more visibility and that can be totally integrate into customer’s system. Incoming data flow Be able to integrate all kind of sources Metadata model flexible (no predefined data model) Outgoing data flow Open/push it Be able to integrate all kind of sources Flexible metadata model SEO compliant Google Google Scholar
• #12: Last but not least, once your data are cleaned and structured, you are ready to use a very powerful « Analytics Dashboard » Productivity metrics & Impact metrics Collaboration & reference graphs Data visualization for strategic intelligence The system can also export report with customized/pre-defined cover. With your complete database we will be allow to create your own analytics reports, for example: Average number of funded studies by country, by type of funding, by type of collaborations Trends of research projects topics (by country, type of funding…) Fields of the studies (based on your “research categories”)