Roadmap
The AODN portal is a web application for finding and downloading geospatial data. We are making changes that will enable you to download data in three easy steps:
- You will find datasets by browsing classifications (such as parameter, location and organisation) in any order and in any combination.
- You will be able to refine and visualise the datasets that you have selected. (eg. filter, aggregate, animate, and plot).
- You will download your data with a single click. The download button will be clear and obvious.
The user interface will be responsive, robust, and it will hide backend complexity from the user.
See further down for a mockup of the new interface
This feature will allow users to add data collections to the map for viewing, filtering, and download. The contents of the portal map will BE the contents of the download cart. In other words, the portal map will be a visualisation of the data to be downloaded. A list view will also be available. There will be a prominent download button on the map view. The data can be in three forms:
- Non-gridded data (stored in a DB and served through WMS/WFS).
- Gridded data (stored as NetCDF files and served through ncWMS and the AO-DAAC).
- Additional data attached to the metadata (such as PDF and Excel files).
Non-gridded data (such as time-series and profile data) will be available as either CSV, NetCDF, or KML. Gridded data will be available as either NetCDF, HDF, ASCII, or as a list of OPeNDAP URLs.
Once users have added WMS representations of data collections to the map, they may want to download the data. Users, however, will often not want to download all the data that is represented by a layer on the portal. They will want to subset it first. There are two types of data (gridded, and non-gridded) and these will be treated differently.
Gridded Data
The AO-DAAC application is an application developed by the CSIRO that subsets and aggregates gridded data, split across multiple files, into a single file for download. It is a standalone application, but a tab will be added to the portal's right hand panel that will provide a graphical front end to the AO-DAAC API. This will be integrated with the enhancements to the download and animation features.
In the case of IMOS at least, gridded data is served through ncWMS. Behind the scenes it is stored in NetCDF files. ncWMS allows the data to be visualised and animated in the portal. Subsetting is made possible by the AODAAC. The AODAAC also allows subsetted data that spans more than one file to be aggregated into a single file and returned to the user. In other words, the user can specify a space/time region to be visualised through animation, and then downloaded as a single netCDF or HDF file containing data only from the space/time region of interest. Both the animation and aggregation features already exist, but these will be merged into a single feature with a single user interface.
Non-gridded Data
In the case of IMOS, non-gridded data is again stored as NetCDF files. This data will be harvested into a relational database and served through GeoServer as WMS layers. GeoServer also supports WFS. Using CQL queries sent to GeoServer, the portal will allow the user to visually filter non-gridded data on the map and then download the underlying data from the database. The data will be downloadable in a variety of data formats (see next section).
In the case of Argo data, there will be two different views. The first one will represent the last location of the floats on the map. The other view will contain all the data for all the parameters and all the profiles. A mapping between the two views will be necessary as when you are filtering the map you will want to download the corresponding data. This type of configuration will be needed for most IMOS datasets (profile data like Argo or XBT and also timeseries like moorings or ships data)
Portal users find it difficult to get the data they are interested in from the portal. Part of the problem is the difficulty in finding the data. Finding data is particularly difficult in the AODN portal because the menu tree is so large, and because it’s ordered at the top level by organisation. Most users don’t care about the organisation and find it an annoying place to start. Some datasets (such as ALA) are so large that they cannot even be added to the menu tree. The faceted search will help solve these problems. The faceted search will give the user more opportunity to examine the layers (not only by organisation). Several views will be presented at the same time, the user will be able to refine and expand the different searching elements.
The term "faceted search" is a fancy term for a navigation system seen on many web sites. It allows users to refine a list. It can be a list of anything at all. It can, for example, be a list of cars. The faceted search allows the list to be refined by multiple classifications - e.g. make, model, colour, price, location, etc. Eg. carsales.com.au
The AODN hosts a catalog of data collections. The faceted search will allow the AODN’s list of data collections to be refined in the same way that carsales’ list of cars can be refined. The AODN data collections will be refinable by classifications such as parameter, location, temporal extent, platform, organisation, etc.
Once the data collections have been narrowed down using the faceted search, they can be selected and added to the portal map. In this way, the faceted search can replace the menu tree. It will also replace the current portal search, which is just a different front end to the same function.
Rather than a line of text (as shown above) individual search results in the list could be formatted to look like a carsales search result. Perhaps something like this:
The view button has the dual purpose of adding to the map/cart. If the result is already on the map the add symbol will be removed and it will look like this:
At a minimum, the facets (ie the classifications by which the list can be filtered) will include:
- Parameter
- Location/Bounding Box
- Temporal range
- Platform
- Organisation
- Free text
- Instrument
- Taxonomy
After performing a search, a facet count will appear next to each values. It will show how many results match each value.
The right hand panel could contain some additional features, such as the ability to save searches. See the following website for some ideas: (http://www.gopubmed.com/web/gopubmed/1?WEB0zh2vqiedyju5I2fI4pI00h001000d10020000011000&rl )
A breadcrumb trail should be included because it shows what constraints have already been applied and allows for their removal. There are different ways of doing it. In carsales the filter selected appears in the left hand panel. In other websites the selected filters appear at the top of the search results.
The search results will be refined on-the-fly and will show a thumbnail view of the data collection (as depicted above). The results will also show the collection name, the organisation, the parameters, and the temporal extent. Hovering over the collection in the list could show a full abstract on the right. Each search result will also include a “view” button for adding results the map. Clicking on this button will automatically take users to the map.
The results presented on the carsales website are a good example because they contain information corresponding to most of the facets. People get a feeling of how they can narrow their search by using another facet. Information about how to access the data could be added to the search results. For example, for glider data, we could have a link to the Thredds folder where the NetCDF files from all the campaigns are available available.
We will also add a feature to allow the user to sort his or her results. e.g. by date, by popularity, or alphabetically by title. Geonetwork has this functionality.
At the back end it works exactly the same way as the existing portal search. ie. It is a GeoNetwork search coupled with the spatial index.
Search for collections using facets such as parameter, location, platform, organisation etc. Search results appear in the middle.
Visualise and subset selected collections. For example, select data from a particular ship in a particular year - results are reflected in what you see on the map. Data can also be animated, charted etc.
Download the subsetted data (shown on the map) in any format.
