1. Inhalt
  2. Navigation
  3. Weitere Inhalte
  4. Metanavigation
  5. Suche
  6. Choose from 30 Languages

Crowdsourcing

Collective Intelligence: has the golden age of the 'citizen scientist' finally arrived?

After years of skepticism about the contribution ordinary people can make to scientific research, the latest thinking is citizen scientists do good science. But can it be trusted?

A study by IIASA - the Vienna based International Institute for Applied Systems Analysis - was based on crowdsourced data from a Geo-Wiki project in which citizen scientists examine satellite images and classify land use data.

"You first see the start screens, help signs and now you get into an 'about' screen, which explains what the app is about," says Tobias Sturn, a young computer graphics expert who is developing an app to help citizen scientists on the Geo-Wiki project.

In the centre of his tablet screen is a land cover map, a Google satellite image of some Norwegian forest.

"And you now have to decide if there is any crop land on this map - choose No, Yes or Maybe," instructs Sturn. I choose "no" and am rewarded with a message telling me I have validated data for an area the size of the Vatican.

A tractor on a farm

Is crowdsourced data about cropland reliable?

Masses of data vs. crowds of citizen scientists

The app and games designed by Sturn are part of a project involving ordinary people in unusual science - a global survey of land use.

"So citizens can actually help through the use of Google Earth imagery to provide very valuable information in terms of what they see on Google Earth - in terms of land cover and land use," explains Steffen Fritz, a remote-sensing expert from IIASA.

Fritz says the crowdsourced information is then combined with other data from modeling tools and used for modeling outputs.

Citizen Scientists have been around almost as long as science. But two key recent developments - the spread of the Internet and access to global satellite imagery - could mean a bright future, especially for the earth sciences.

Steffen Fritz, a remote-sensing expert from International Institute for Applied Systems Analysis
(Photo: Kerry Skyring)

Steffen Fritz, remote-sensing expert, IIASA

But can the data being fed back to the experts by crowdsourced amateurs be trusted? Fritz and his colleagues at IIASA have compared results from both groups and their conclusion is yes, in many cases it can.

"We were quite surprised actually that ordinary citizens could deliver quite high quality data, so this was actually new," says Fritz. He concedes not all citizen scientists are the same but there are some who perform as well as the experts.

Experts in their own right

Guwahati, a major city in northeastern India is a place of great natural richness and diversity. But it is also a place where land use is rapidly changing.

One of Guwahati's residents, Rubul Hazarika, saw the crowdsourcing tools of Geo-Wiki and decided to do his bit.

"Citizens are experts in their own right," he says, adding that in his experience citizens are specialists in their farms, in their forests and in their fishing communities.

"Scientists may possess the methods and tools for scientific study but citizens hold traditional knowledge," says Hazarika, who teaches geography at a local college.

Screenshot of the Geo-Wiki project game, CroplandCaptcher.
(Photo: CroplandCaptcher/IIASA/University of Freiburg)

The CroplandCaptcher app

To assess the value of research results from people like Hazarika, IIASA gave 60 people - experts and crowdsourced non-experts - more than 50,000 images to analyze.

The experts got it right 69 percent of the time, while the non-experts got it right 62 percent of the time.

They concluded that with a bit of training, the non-experts could match the experts.

Steffen Fritz says this is important work.

"There are these high uncertainties in global land cover," says Fritz. "There are areas especially in developing countries where we don't know where the crop land is for example and because of those high uncertainties we decided to build this crowdsourcing tool called Geo-Wiki."

He says the idea is to have ordinary people view these uncertainties and tell them - the scientists - where the maps are correct and where the maps are wrong.

Every home a digital laboratory

Anna Cipriani, who lives on the outskirts of Vienna, is another citizen scientist who took part in the IIASA data validation tests. She's a qualified scientist in her own right, but volunteered for the project on land use.

Anna Cipriani is a Citizen scientist from Vienna, Austria (Photo: Kerry Skyring)

Anna Cipriani, scientists and 'citizen scientist'

"I find it very exciting because for example the other day my husband and I were watching a documentary on TV and we saw some images of Africa and we thought, Oh, that's what it is…you know, they have these villages with strange fences and we were validating these images from Geo-Wiki!"

IIASA admits some training of citizen scientists will be necessary to get the best results. They are developing tools to do this online.

Anna says it shouldn't be too hard for most people to get to grips with the basic concepts.

"I think that if you could spend a few hours with an expert, do like a hundred or two hundred images…together with the expert, I think you could go on and do it as a job," she says.

The next challenge for citizen scientists

Using Google satellite imagery to check human impact on land cover is a fairly obvious task for crowdsourced scientists.

But Fritz is working on some new challenges, such as engaging local people in collecting data on water reservoirs and dams.

"We know where the big reservoirs are," says Fritz, "we know where the big hydro electric power schemes are, but we don't know where the small ones are and the small ones, especially in developing countries, are very valuable and important."

DW recommends

Audios and videos on the topic