Tuesday, December 12, 2006

A Global Indicator Information System

Indicators measure our progress toward goals, at an individual, regional, and global level. There are innumerable benefits in organizing and sharing this information globally. Provinces in separate continents with similar social demographics and environmental conditions can, for example, share how they have successfully resolved their common problems and improved the conditions that lead to a better quality of life. Records and distributions of indicator data at various administrative levels also educate administrators about their neighbors, as well as giving them wide overviews of issues. However, with all the benefits in a framework of shared indicators, one has yet to be established.

A global indicator information system gathers, analyzes and shares indicator data, bringing it to the widest audience possible. The purpose of this article is to define and discuss the principals and components of such a global indicator information system. These include:

* A protocol to communicate indicator data between the system parts
* Thematic mapping software to visualize the information
* Statistical analysis tool to interpret the data and convert it to information
* Geographical map and satellite imagery service broker and warehouse
* An open-information and open-source system
* Security model to protect data integrity and ownership
* Geocoding standard

The troubling aspect of each system part will be explored, as well as some working solutions and standards.

Indicator Communication Protocol

In order for one component of an indicator information system to communicate its processed data with the other parts of the system, a standard protocol is required. CRIS, the Country Response Information System, a statistical analysis tool developed by UNAIDS, communicates georeferenced indicator data using an XML-based protocol called the Indicator eXchange format (IXF). IXF has a well defined schema that allows for multi-dimensional datasets, complex indicator relationships and source information at the data node level. The protocol has already been adopted by several thematic mapping applications, including KIDS (FAO), DevInfo (UNICEF) and HealthMapper (WHO). These systems can import indicator data from an IXF source, and also share their data in the proposed format.

Thematic Mapping Software

With so many thematic mapping software available currently, any attempt to analyze the features, benefits, and differences of them all, would be futile, as the products will be obsolete or updated before their review is published. Rather, looking at some superlatives of indicator data will be a better measuring stick of the capabilities of thematic mapping software.

Having worked with indicator data from various fields, such as health, agriculture, animal disease, fisheries, administration and environment monitoring, a common shortcoming of most thematic mapping applications has been the treatment of indicator data in limited dimensions. Often, besides the temporal and spatial dimensions, only one other dimension is allowed for the indicator value in thematic mapping applications. Therefore a lot of information is aggregated into one quantity, which would otherwise empower the user if he/she were given the choice to select and control how he/she wanted the aggregation to occur. The following example from KIDS, the Key Indicator Database System developed at FAO, shows how multi-dimensional indicator data can be successfully handled.

Another crucial test for a thematic mapping candidate is its performance in rendering high definition maps. Rendering maps with millions of polygons and points can easily bring a system to its knees. Robust and well crafted thematic mapping software will respond in real-time to user zooms and scrolls, producing images with just enough resolution and depth.

Other factors to consider in selecting the thematic mapping application are how well it handles metadata and what data attributes it supports. Attributes such as the precision, units, source and references of the data should be available. The indicator legend may be predefined, but if not, it is the responsibility of the mapping software to dynamically generate the legend based on the data distribution curve.

The software should also support multiple projections, which may not be as interesting for maps at district and province levels, but become significant in importance from country, continental and global perspectives.

Geographic Warehouse and Service Broker

Maps are an essential component of an indicator information system. Analysts, decision makers and GIS experts need access to digital maps to spatially view their indicator data. To avoid duplication and increase both cooperation and the quality of collected data and maps to be made available to everyone, a distributed network of map databases is necessary. In addition, to search for and find the desired map, a smart query system based on the map’s metadata needs to be implemented. Some protocols and standards have already been suggested for such a global network of map sources and the searchable metadata, by the OGC. One successful implementation of these protocols, along with an open source application for running your own map warehouse and query client, is GeoNetworks.

Statistical Analysis System

Data need to be processed before they can be mapped. Often, this requires filtering the data through a statistical analyzer. For example, AIDS cases in Botswana alone do not have much meaning unless they are disaggregated over age groups and gender, then given percentage values of prevalence over the whole population of the country, and compared across the various districts. Or there may also be too much census data, and a statistical system can aggregate that information into what can be easily visualized on a map, chart or time-series graph.

Open Information/Open Source

Information is vital to developing countries in combating AIDS, poverty and hunger, and to ensure education, environmental sustainability and social equality. Most NGOs, UN organizations and funding bodies support programs to gather data and disseminate it in an informative way that assists development where development is needed most. And most agree that there should be no price tags or restrictions on data that can be used to educate the policy makers to make informed decisions on how to best invest in the social welfare of their jurisdiction. But is proprietary, copyright protected software a means to circumvent the unethical practice of information hiding?

The open source model is an essential principal in building a global indicator information system. There is no business or corporation in the world that can afford to address all the needs of all the clients who use software to analyze indicators. The only model which can successfully support such a massive demand is an open source community. Therefore, it is absolutely essential for any global indicator information system to be armed with a toolkit of open source software and open-information services.

Security Model

Data are precious and proprietary. Data collectors want to control how they share their data and want exclusive rights to who can modify it or see it. KIDS implements such a security policy, where features necessary to view, modify and download data are made available to clients based on a tiered set of access controls. Any other system connecting to the KIDS Web services to collect data must authenticate itself and will be bound by the rules and permissions given to it. Thus data integrity and security is preserved, even in an open source, open information system.

Geocoding Standard

Conflict, politics, and policy in many developing and under-developed regions change the face of our world daily. Administrative boundaries shift, district names change, and provinces merge or split to form new ones. The implication and challenge for GIS is to keep the maps and data updated constantly and accurately. To ensure accuracy and avoid the tedious and manual task of matching data to maps, standards need to be devised and conformed with. So a well devised standard should address the georeferenced data and boundary map units with “geocodes” that identify the administrative level of the region, the region’s ISO 3166-2 code, and the time when that data was recorded or the region was coded.

An initiative called GAUL (Global Administrative Unit Layers) has been launched by the Department of Sustainable Development, which addresses the issues surrounding geocoding of boundary maps. Its principles have been well received by the many UN groups. Its success however, remains dependent on its wider acceptance and adoption by the larger GIS community.