Crowd-Sourced Database for Biomaterials

Client

Ashim K. Datta, Professor of Biological and Environmental Engineering
<akd1@cornell.edu>

The client is very interested in this project. Process simulation is his research interest and properties data have been his interest for 25+ years. He is very passionate about building this database and would be willing to meet with students as often as needed.

Student contact

This project has an almost full team. If you are interested in joining the team, please contact Nathaniel Leo Jones <nlj24@cornell.edu>.

Background

Physical properties data are key components in 1) understanding materials and processes; and 2) performing engineering simulations of processes. Currently, such data are buried into graphs and tables in individual research papers—they are not searchable, importable into engineering software and are not amenable to seeing the trends (critical in material selection). In some cases, this is effectively the same as if the data never existed.

This problem is generic, i.e., it exists for all materials. For engineering materials, there are some databases available but, for biomaterials, almost nothing is available. As an application example, our group develops engineering simulations of food processes such as microwave heating of food that involves thermal and dielectric properties (thermal conductivity, specific heat, dielectric constant, dielectric loss). These properties change with temperature (as the food heats) and moisture (as it loses moisture) and of course they depend on composition. In the area of food materials, the user base and the research funding are small enough that although it will be tremendously useful in industry and academia, it is unlikely to become a commercial venture. Crowd-sourced building is about the only possibility.

The most extensive such database available for non-biomaterials is at http://www.grantadesign.com/. It is very expensive and is for profit and not crowd-sourced. I do not know how much it is used but it is very unlikely to be a particularly popular one. The best example of a crowd-sourced database that has similarities in type of information and user base as the one being proposed here is in the Chemistry community (http://www.chemspider.com/). I have direct contact with the creator and he is very interested in helping us if we were to go forward (but the properties we are discussing will not fit in that database).

Goals and Deliverables

To develop the interface and the structure for a crowd-sourced physical properties database that can take measured data as well as correlation equations that can predict data (e.g., predicted as function of temperature and composition). All prediction equations have been already compiled by our group and are ready to go. We also know the sources of raw data and only enough of this data would be entered to test the functionality of the software (i.e., project is not about exhaustive data entry). Anyone with web access should be able to contribute to the database (so that it grows and is sustainable) but there needs to be a gatekeeper for minimal checking of accuracy. The interface needs to have visualization capabilities that would include measured and predicted data together and we should have buttons for choosing between ranges of parameter values (much like how we have buttons for choosing between a price range or a capability in a store website such as the one for Best Buy). As said earlier, the concept is generic so it will be applicable to materials outside of the specific materials chosen for this project.