J14.3
Multigraph: interactive data graphs on the web

- Indicates paper has been withdrawn from meeting
- Indicates an Award Winner
Wednesday, 20 January 2010: 9:00 AM
B218 (GWCC)
Mark Phillips, Univ. of North Carolina, Asheville, NC; and D. Eldreth

Presentation PDF (688.3 kB)

The fields of weather and climate are full of time-series data. Observations, model output, forecasts --- almost all aspects of weather and climate involve time dependent data that is often presented in the form of a graph. With this in mind, and considering that the scientific community and the world at large has been turning for several years now to the web as a primary means of communication, there are surprisingly few good tools and techniques available for presenting time-series data on the web. The most common solution is to use a desktop tool such as Excel or Matlab to create a graph which is saved as an image and then included in a web page like any other image. Although this is straightforward and simple, it limits the user to one particular view of the data, disconnects the graph from the data in a way that makes updating a graph with new data an often cumbersome manual process, and does not provide the reader with the ability to combine and compare data from separate sources.

This situation is somewhat analogous to the state of mapping before the advent of GIS. Maps existed only in printed form, and creating a map was a laborious process. There were many disparate sources of maps, and information from separate maps could only be combined manually by creating a new map. In the last several years, however, the world of mapping has experienced a revolution in the form of web-based and other interactive computer technologies, to the point that it is now commonplace, for example, for everyone from small children to great grandparents to easily browse through gigabytes of geographic data to locate an aerial view of their house or neighborhood, or create and print a customized trip map complete with driving directions and time estimates.

Multigraph seeks to bring a similar ease of presentation and viewing to time series data. Multigraph is a program for displaying interactive time-series data graphs in web pages that includes a simple way of configuring the appearance of the graph and the data to be included. It allows multiple data sources to be combined into a single graph, and allows the user to explore the data interactively. Multigraph lets users explore and visualize "data space" in the same way that interactive mapping applications such as Google Maps facilitate exploring and visualizing geography.

Viewing a Multigraph graph is extremely simple and intuitive, and requires no instructions. Creating a new graph for inclusion in a web page involves writing a simple XML configuration file and requires no programming. Multigraph can read data in a variety of formats, and can display data from a web service, allowing users to "surf" through large datasets, downloading only those the parts of the data that are needed for display.

Multigraph was created at UNC-Asheville's National Environmental Modeling and Analysis Center, with support from the Renaissance Computing Institute and NOAA's NCDC, and is freely distributed as open source software. It has been deployed on several NOAA web sites including the US Drought Portal (http://www.drought.gov), the Climate Reference Network (http://www.ncdc.noaa.gov/crn), and NCDC's Climate Data Online (http://www7.ncdc.noaa.gov/CDO/cdo).

Use of Multigraph in other contexts is encouraged. It can be downloaded from http://www.multigraph.org. The web site also contains documentation, examples, and information on how to join the Multigraph Users mailing list.

Supplementary URL: http://www.multigraph.org/ams2010paper