Today’s featured dataset is the Global Hunger Index. The GHI set includes information about global nutritional poverty linked with geographic identifiers (Geonames). It’s been contributed by Tim Davies, who kindly answered a few questions I had about the dataset. I’m keen to see humanitarian datasets can be put to use, and would love it if you were to see some potential here!
What is the Global Hunger Index, and what data is in the set?
The Global Hunger Index (GHI) is designed to comprehensively measure and track hunger globally and by country and region.
It is calculated each year by the International Food Policy Research Institute (IFPRI) and it highlights successes and failures in hunger reduction and provides insights into the drivers of hunger. By raising awareness and understanding of regional and country differences in hunger, the GHI aims to trigger actions to reduce hunger.
On the Kasabi platform we’ve now got two years of the GHI Dataset, giving scores for over 100 countries for five different years. The different years of the GHI data aren’t directly comparable, so the 2011 release of data updates 1990 figures for the 2010 release. The data is linked to Geoname identities for each country. You can find both the GHI figure for countries and years, and the supporting data used to calculate it (see below).
How do you see the data being used?
The data lends itself to being used in existing country profile pages as a key statistic. For example, the Food and Agriculture Organisation of the UN have integrated the data into their country profiles.
The data is also good for use in mapping mash-ups. So far we’ve used the tabular data to create an interactive map and others have used the data to create maps like this one on ChartsBin, but with linked data there’s the potential to pull in a lot more context about each county – from population numbers, to other key statistics or profiles.
We hope as well that introducing this data into the Linked Open Data Cloud will raise awareness of it – and that users will come up with creative ways to draw upon it.
What is the structure of the linked data?
It’s really a very simple dataset.
All the statistics are in the store as single measure RDF Data Cubes. There’s a dataset for each yearly release of the Global Hunger Index data (so there’s a 2010, and a 2011 dataset in there) to help users identify the data that is comparable (only data from the same dataset can be directly compared). Country and measure are key dimensions of each datacube.
We’ve set up a few example queries to show how to fetch all the values and supporting data for a particular country, or from a particular GHI release.
All the URIs are in the space, which means all the data can be directly dereferenced as linked data.
In our first pilot putting together this data we experimented with a number of different formats, all documented at http://data.ifpri.org/rdf/ghi/ but found that the Data Cube version was used more than a version using the SCOVO vocabulary.
We’ve also documented the process used this year to transform the raw GHI spreadsheets into linked data. We used Google Refine with the RDF and Kasabi extensions to upload our data direct onto the Kasabi platform which made the whole process a lot easier.
How is the Global Hunger Index put together?
A number of different indicators can be used to measure hunger. To reflect the multidimensional nature of hunger, the GHI combines three equally weighted indicators in one index number:
- Undernourishment: the proportion of undernourished as a percentage of the population (reflecting the share of the population with insufficient calorie intake);
- Child underweight: the proportion of children younger than the age of five who are underweight (low weight for age reflecting wasting, stunted growth, or both), which is one indicator of child undernutrition; and
- Child mortality: the mortality rate of children younger than the age of five (partially reflecting the fatal synergy of inadequate dietary intake and unhealthy environments).
This multidimensional approach offers several advantages. It takes into account the nutrition situation not only of the population as a whole, but also of a physiologically vulnerable group – children – for whom a lack of nutrients creates a high risk of illness, poor physical and cognitive development, and death. In addition, by combining independently
measured indicators, it reduces the effects of random measurement errors.
The GHI ranks countries on a 100-point scale. Zero is the best score (no hunger), and 100 is the worst, although neither of these extremes is reached in practice. The scale on the following page shows the severity of hunger – from “low” to “extremely alarming” – associated with the range of possible GHI scores. The 2011 GHI is calculated for 122 countries for which data on the three components are available and for which measuring hunger is considered most relevant (some higher-income countries are excluded from the GHI calculation because the prevalence of hunger is very low).
The GHI is only as current as the data for its three component indicators. This year’s GHI reflects data from 2004 to 2009 – the most recent available country-level data on the three GHI components. It is thus a snapshot not of the present, but of the recent past. For some countries, such as Afghanistan, Iraq, Papua New Guinea, and Somalia, insufficient data are available to calculate any value for the GHI at all. Even though abundant technological tools exist to collect and assess data almost instantaneously, enormous time lags persist in reporting vital statistics on hunger. More up-to-date and extensive country data on hunger are urgently needed – a situation explicitly recognized by the Group of Twenty (G20) countries in their 2011 action plan on food price volatility (G20 2011). Improvements in collecting high-quality data on hunger and food consumption will allow for a more complete and current assessment of the state of global hunger and, in turn, more effective steps to reduce hunger.
The source data on which the GHI scores are based are continually revised by the United Nations agencies responsible for compiling them, and each year’s GHI report reflects these revisions. The revisions result in improvements in the data, but they also mean that the GHI scores from different years’ GHI reports are not comparable with one another. This 2011 report, however, offers an advantage over other recent GHI reports in that it contains not only the 2011 and
1990 GHI, but also GHI scores for two other reference periods – 1996 and 2001 – that are comparable with one another, allowing for in-depth analyses of trends. In other words, comparable source data were used to calculate the GHI scores for all four reference periods in this report.
More information at http://en.wikipedia.org/wiki/GlobalHungerIndex