| Project Date -- 2018
| Link to Project -- https://youtu.be/hVDZtmFRfIQ
| Repository -- https://github.com/liviajakob/data-sharing-platform
Project Overview | IceExplorer - Building a Visual Data Sharing Platform
The IceExplorer web interface can be accessed through the following URL (only accessible from the University of Edinburgh network):
The IceExplorer code can be found in the following repository:
- github.com/liviajakob/data-sharing-platform (tag: v1.0)
The IceExplorer demo video can be found here:
The open access movement, promoting research articles to be freely accessible, has expanded to include the data associated with the research, however, there is a lack of platforms designed for research groups to share their data. In addition, there is little knowledge about researchers' expectations and preferences for data sharing tools. This research is important as platforms fitting the specific needs of research groups could encourage them to share their data. This project aims to address these gaps by investigating the requirements of research groups for geographic data sharing tools and exploring techniques to visualise and contextualise spatial datasets within such a tool.
As a proof of concept, the IceExplorer - a prototype of a web based visual data sharing platform - was implemented, containing Earth Observation data. Components include (1) a Graphical User Interface (GUI) with spatial visualisation and the functionality to download data, (2) a web Application Programming Interface (API) to download data directly and (3) a tool for researchers to add new data via a command-line utility.
A User-centred Design (UCD) approach was applied, including context of use analysis, user requirement survey, user requirement specification, user usability testing and user survey. The findings during the UCD process have confirmed the importance of visualisation for spatial data sharing. Visual data sharing platforms - such as the IceExplorer - motivate researchers and research groups to share their data and support data users to easily access desired data products.
Keywords: Data Sharing, User-Centred Design, Data Visualisation, Web Application, Earth Observation Data
User-Centred Design Approach
A User-centred Design approach was applied, including context of use analysis, user requirement survey, user requirement specification, user usability testing and user survey. Below a diagram of the User-centred Design workflow used in this project.
The following main tools and technologies were used:
- python3 -- server-side
- Flask -- python web framework
- JSON and GeoJSON -- data exchange
- GDAL -- for raster manipulation (gdal2tiles for generating tiles)
- SQLAlchemy -- SQL toolkit to make database flexible (also use of ORM)
- SQLite -- database for meta data
For more information see: github.com/liviajakob/data-sharing-platform
Server-side -- Command-line Tool
A command-line interface offers the option to the research group to register new data which is then processed for visual display and ingested into the system. The IceExplorer command-line utility provides two endpoints; one to ingest a new dataset and one to add a layer to an existing dataset.
Server-side -- API
The RESTful web API of the IceExplorer provides the option to directly query and download data without using the GUI. Two different endpoints are available; (1) query available datasets and layers and (2) download the data itself. A typical API request URL of the first endpoint could take the following form:
Client-side -- Web GUI
The web-based Graphical User Interface offers two views: the Data Catalogue View and the Data Explore View.
The Data Catalogue View - the entry point of the application - is logically linked to the side panel and shows the geographic extent of the datasets listed in the side panel on the map as polygons. It serves the purpose to give the user an overview of the available data, however, the data itself cannot be accessed within this view. The map is interactive and provides controls to zoom, show scale, show an overview map, display coordinates of the mouse position, toggle fullscreen, refresh map and switch background layer. The user can currently filter datasets by data type (DEM, rate, error, velocity and radar backscatter) and time span.
The Data Explore View is activated when the user clicks on one dataset to view the actual data. Within this view, a detailed infobox with dataset information and the available layers to download and switch between are displayed.
In addition, the Data Explore View provides a toolbox with three tools; the Swipe Tool to compare overlapping data, the Get Values Tool to query values over time by clicking on a position on the map and the Interactive Legend showing the values of the colours of the top visible layer.
To conclude, it can be stated that this project has successfully built a prototype of a visual data sharing platform, including (1) a command-line interface for data providers to enter new data, (2) a Graphical User Interface to visualise and download the data and (3) a web API where more experienced user can access the data. The technical implementation of the IceExplorer uses open source projects and comprises of a flexible design to facilitate further development.
Furthermore, User-centred Design has proven to be particularly useful to bypass the challenges of visualising geographical data in a comprehensible way and is recommended as a technique for any future GIS application development. However, future research is needed to refine the use of the UCD process in a GIS context to identify the best practices and establish standards.
Finally, the findings and perceptions during the UCD process have confirmed the importance of visualisation for spatial data sharing. Visual data sharing platforms - such as the IceExplorer - motivate researchers and research groups to share their data and supports data users to easily access desired data products.