Efficient Data Management at CPI

• Devin Cummings, Computer Science, University of Delaware

• Hui Fang, Department of Electrical and Computer Engineering, University of Delaware

The current Center for Plastics Innovation database of research articles is not the most efficient for downloading articles, as viewers have to go through a 3rd party website where they often have to login or even pay to download articles. What we created was a more streamlined site where articles can be downloaded straight from one location. In order to accomplish this, we utilized three coding tools. First, Elasticsearch allows for the publishing and retrieval of articles from a centralized database. Kibana was then implemented to better visualize Elasticsearch. Lastly, Python Flask was used to publish and run the website. Presently, the site has three main screens. Whenever a user attempts to access the site without being logged in, they are redirected to a login screen. Here, they can login with whatever credentials they have, and if they don’t have any they have the option of creating a new login or signing in with Google. On the search screen, users can search for articles based on keywords in the file name, content of the file, or by column if the file is an excel type, and there are buttons to download each resulting article. Additionally, there is the ability to add new articles after clicking the “add new file” button on the search screen. On this page, files can be uploaded both individually and in bulk, and each of them have an associated file type (PDF, excel, or general type), a description, and a label for what CPI thrust it pertains to. Web development is a fluid process, so we hope to add more features in the future, like adding a diagram to illustrate the connections of what authors have worked together, creating an admin login which can delete files from the database, and general improvements in user friendliness.