It’s been a while since I’ve worked on this site, but thanks to the Food+Tech Hackathon I’ve taken the time to really think about how the data and tools I’m building should evolve. If you have any questions, comments, or suggestions, please feel free to email me at ewpatton [at] gmail [dot] com so we can begin a dialogue.
Firstly, I’ve uploaded and made available all of the data I’ve processed from the United States Department of Agriculture. This includes their Nutrient Database for Standard Reference and recommended dietary guidelines. Dominic DiFranzo and I have also assembled a recipe ontology for describing recipes that take advantage of these data sets. I’m in the process of documenting the ontology in human-readable format (daring souls can perform direct queries for labels on properties and classes), so please be patient as I work on generating these documents.
Also, I would like to thank everyone who participated in the hackathon, especially those individuals who took advantage of either the semanticdiet.com data or the Wine Agent’s food ontology. I was involved in many great discussions with participants (esp. Danielle, Britta, Sean, and Tian) and I hope that further collaboration between us all will lead to better data for the benefit of everyone. In the meantime, I will continue to be exploring datasets and thinking about interesting future steps/projects to take advantage of this wealth of data.
All of the data files are available both as raw RDF files and via a live SPARQL endpoint for online query. If you are not familiar with the SPARQL query language, you may want to check out the SPARQL Query Language recommendation from the World Wide Web Consortium or the Jena Framework’s tutorial on SPARQL queries. The current datasets do not contain the provenance information, but an updated dataset will be made available in the next few weeks that tracks where the data come from and the transformations I’m applying to generate the structured data.
For those of you who might have tried using the two ontologies together, you probably noticed that the nutrient data and the food ontology are not yet linked together. This mapping needs to be established from the current data, especially linking the more relaxed nutrient ontology with the more expressive food ontology. There are some current incompatibilities between the two datasets that need to be resolved, and I also want to look for an automated approach to linking the two ontologies together.
Originally posted at http://blog.semanticdiet.com/
Evan Patton is a Computer Science Doctoral student at Rensselaer Polytechnic Institute in theTetherless World Constellation. Hi primary focus is in adapting Semantic Web technologies for use on mobile devices as well as parallelization of reasoning techniques to take advantage of many heterogeneous machines for the purpose of solving complex tasks, with the Tetherless World Wine Agent as a test bed application of this work.