Wiki

Clone wiki

gnd / DataIntro

Sample data

Here's how to find some sample data.

The data is in a hosted CouchDb app at: http://gnd.iriscouch.com/_utils/

The sample data is in the tracks database, at: http://gnd.iriscouch.com/_utils/database.html?tracks It's a flat collection of JSON documents.

Here's a readable listing of tracks: http://gnd.iriscouch.com/tracks/_design/tracks/_view/track_listing

The id field in that JSON document is the database index. So, you can retrieve a document using: http://gnd.iriscouch.com/tracks/0d9500c05ef3a901397b5d7dda122de4

There's also a view that shows just the metadata for tracks: http://gnd.iriscouch.com/tracks/_design/tracks/_view/metadata (it's this view that I intend using for the ElasticSearch Couchdb river - as described in the plugin at: https://github.com/dadoonet/elasticsearch-river-couchdb/tree/dpi_views#using-couchdb-views)

And here's a pretty view of a document to show the file structure: https://bitbucket.org/ianmayo/gnd/wiki/GND_File_Format#!structure

Size of data

The sample database above currently contains around 500 documents. I expect the storage to grow at approx 1000 documents per month.

Response time

I expect ElasticSearch queries to appear instant to the user, so I guess around 1/2 second would be acceptable.

My Web and Desktop interfaces will use faceted search. This UI mockup shows drop-down lists of facets: https://gomockingbird.com/mockingbird/#nq36az7/9zEEWl

For these fields I wish to allow the user to select the term from a list, plus I wish the contents to be covered by a free-text search:

  • platform, platform-type, sensor, sensor-type, trial, type For these fields I wish to allow the user to specify time limits (before, after, between):
  • time_bounds

Search UI

I'm torn between a faceted search, and the power of free-text search. Clearly the second UI shown below is intuitive: simplicity.png

But, the terms do have a logical grouping. So I suspect the most effective way is to have free-text as the top-term, then restrict the facets based on the results of the free text search: images/FacetSearch_UI.png

Updated