In-Depth

Case Study: IBM, Predictive Analytics, and Grevy's Zebra

With the world losing one or more entire species of animal or plant life every 20 minutes, a conservation group in the UK is using analytical software from IBM to untangle the complex factors contributing to a particular species' decline.

• By Linda L. Briggs
• 12/01/2010

IBM's iconic striped logo might not immediately bring a zebra to mind, but for an endangered species now relegated to small areas of northeastern Africa, that is just what it stands for. Grevy's zebra is being studied by the conservation group Marwell Wildlife using IBM's SPSS analytics software. The animal, which has a classic zebra appearance, used to live across much of Africa. After catastrophic declines in its population in the 1980s and 1990s, fewer than 2,500 of the animals remain in the wild, relegated to northern Kenya and small areas of southern Ethiopia.

To study the zebra's decline in the hopes of reversing it, Marwell, a UK-based charity dedicated to the conservation of a wide variety of wildlife and natural habitats, especially in Africa, is using IBM's predictive analytic software to analyze information collected in the field by scientists. Along with Marwell, the project is a collaborative effort including the Grevy's Zebra Trust, the Lewa Wildlife Conservancy, the St. Louis Zoo, and the Denver Zoo.

Predictive analytics software is helping scientists determine where to focus resources in order to conserve the species, while taking into account the needs of people and wildlife. The northern rangelands of Kenya house an array of endangered and migratory wildlife, not just the Grevy's zebra. African wild dogs and elephants also live there alongside human communities. Thanks to efforts including those by Marwell, the decline of the zebra population in Kenya is now being successfully reversed for the first time in over a decade.

According to Dr. Guy Parker, who heads biodiversity management at Marwell Wildlife and is helping spearhead the project, regional conflicts over many years, plus hunting of the zebra and friction over water and grazing land for domesticated livestock, have tended to push out the animal. Its close proximity to human settlements poses threats as it competes with livestock for sparse water and other resources. Information about the species is difficult to collect because its range is so remote and dangerous, Parker says, with bandits, cattle rustling, and illegal weapons in the area.

To counter all of this, Parker's group began collecting data in a study earlier this year to find out how many zebra -- along with other wildlife -- were surviving in a vast area of perhaps 10,000 kilometers.

Using SPSS software, Parker explains, scientists are beginning to analyze the information collected in the field. The data collection process focused on speaking with local people across the region, who are pastoralists grazing sheep and cattle. To collect the data, Parker's group traveled as a convoy in a loop through Kenya's far northern area, speaking with both locals and politicians.

The questionnaire, completed on paper because of glitches with the handheld PDAs the group originally intended to use, focused on what wildlife had been sighted in the area and its status; each of 40 questions had many possible answers. "It might be a small set of data in a commercial sense," Parker says, "but it's a big set of data for us."

From paper, the answers were entered into spreadsheets back at the central office, then imported into SPSS for analysis and manipulation. The many variables set up in SPSS, Parker says, included how hunting related to the number of wildlife, how many hunters were in each area, and locations where different species of wildlife should have persisted but didn't.

In manipulating the data using SPSS, Parker explains, "we also did some quite complex regressions [in which] we looked at multivariate [statistics] and predictives." The data revealed people's attitudes toward wildlife, especially the zebra – attitudes that are hugely important in conservation efforts. In an encouraging result, comparing scores on socio-economic factors such as education, sex, and where a respondent was from, the survey revealed that many people found Grevy's zebra important.

"Normally, when you have people living alongside wildlife, there tends to be some sort of conflict," Parker explains, as with elephants in some areas of Africa. "In the case of the zebra, people see lots of socio-economic benefits." Among them: the animal is a potential source of income from tourism, and in times of drought, it can lead people to good pasture and water. People also revealed their belief in an innate benefit of the zebra -- they are seen as something beautiful in the landscape.

That fact, Parker explains, is important because it makes it much easier to teach conservation principals, and to convince the local population to conserve the zebra. Those insights -- along with major threats facing the zebra -- came directly from the research.

Despite raising sensitive issues such as hunting of the zebra by locals, respondents were straightforward with the scientists in answering questions, Parker says. One interesting data item that the survey and subsequent analysis uncovered was that one of the threats facing the zebra is that it is sometimes hunted for its body fats, which are used for traditional medicines. Conservationists hope that by eventually bringing access to modern medicine to areas, hunting can be curtailed.

Setting up the regressions and other uses of SPSS was done internally, Parker says, with some support from Marwell's IBM account manager. "We mostly did it ourselves; it's very straightforward." Parker says of the software. As is common with scientists, he had experience in using SPSS from previous projects, although this was the first time Marwell had used it. "[SPSS] is very well-suited to questionnaire information and has quite a following in the conservation field," Parker explains.

As the effort to save the zebra continues, data from the survey will be shared with local governments, partner organizations, and survey participants in order to build a picture for the path forward. Plans might include introducing a mobile health clinic to reduce hunting of the zebra for medicinal reasons, or helping communities explore the possibilities for tourism. "We're really just scratching the surface," Parker says, adding that he'd like to start integrating aerial survey information into the database, along with collecting and integrating data from remote camera traps. Over time, the data will allow the limited conservation resources to be used in the most effective ways to counter the threats facing the species.