I have a media set of .png files and a coords.csv holding latitude/longitude pairs for each of the images. I simply turned the media set into a tabular dataset and joined the two datasets together. I made a quick google search, but I’m not exactly sure if there are better formats for my use case.
Here’s my pipeline:
Hi!
That sounds like a reasonable approach to me 
You may also be interested in the Ontology GeoPoint type, that can be created using this transform in Pipeline Builder.
If you use your new table as a backing dataset for an object type in the ontology, you can have add the column with the information about your image as a media reference under the “Capabilities” tab in the ontology manager application. This will allow you to display the image easily in frontend applications such as Workshop.
-Eirik
