First Let'sDig In
We started by debating modeling approaches and considering leveraging the available NASA taxonomies. Since they were older taxonomies, we started building different ontology models. We found keywords and themes within the main NASA JSON dataset, matching the keywords and tags in the NASA Open Data Search portal. However, the portal didn't rely solely on these fields and used more advanced methods. This meant going back to using the available taxonomies to properly categorize the datasets (which, this in itself, was not enough). We tried connecting various categories within the taxonomy, and considered using other metadata fields and combinations of them (i.e. Publication Date, Dataset Collection, Location and even the Data Dictionary of those datasets) .
|
Lights, CamerAction
We modeled those relationships, but for the sake of time, primarily used the taxonomy category connection discovery approach and also made possible connections to the Science categories of the NASA SMD.
We produced different potential ways to model relationships between datasets. Instead of just modeling individual relationships between datasets, the search results can be presented as a visualization of connected datasets with relationships, (as opposed to of a list). Now, connections between all datasets on the NASA Open Data Portal can be added and testing can be carried out via the search portal. |
Join THE FUTURERESULTS
In the future, parts of the different ontology models can be combined to depict both the formal taxonomy categorization as well as the relationship models between the datasets. Ideas from the ‘Discovering Science Connections’ challenge can be applied here, and vice versa.
If given the opportunity and resources, more advanced technologies such as various approaches in AI, may help create the following vision. Imagine an interactive dynamic graphical system that serves as a one-stop-shop for the public to search, find, retrieve, visualize and learn about various datasets from distinct (but possibly related) projects and NASA missions. The user can see information about each dataset, and how they may be associated with others. |