B
N
search for HITs
containing the words "Mapping Africa"
our understanding of where people are farming is limited.
We have a pretty good idea of farmland distribution in Europe and North America, but we have a lot to learn about other parts of the world.
Of particular note is Africa, which is predicted to experience an explosion in agriculture in the coming decades.
The best data that we have so far are not incredibly accurate.
They are prone to overestimating and underestimating farmland in various locations.
Due to the wide range of error and the unreliability of the data, it can be difficult to understand key issues such as food security, or to predict where agricultural expansion will happen.
As a result, we have launched this mapping initiative to get a better idea of where and how much farmland there is in Africa.
In order for our project to be a success, we need you.
Although there are computer algorithms to map these fields, they aren’t as good
as the human eye.
To help us, please visit
Once your account
has been registered by Amazon,
You are ready to start mapping
The process works like this:
Choose to view and accept HITs from the requester “Mapping Africa”.
Identify crop fields falling totally or partially
within the white box in the center of the satellite image.
Fields usually have a semi-regular shape that is distinct from their surroundings,
Fields often have traces of parallel lines (because of ploughing) inside their boundaries.
If there are any, mark each field by drawing a polygon along its boundaries.
If the field falls partly outside of the box, please also map the part that is outside.
Repeat this procedure for all fields within the box.
Submit the HIT.
Start another one
We will pay you for each HIT, assuming that your quality score receives at least 60 out of 100 possible points.
How do we determine your quality score ?
Well, every once in a while we sneak in a HIT we have already mapped, which allows us to measure how well your map aligns with ours.
We assess your work's quality according to two components:
So the process and scoring is fairly simple.
However, interpreting what is and isn’t a field can take some practice, and we are also only interested in certain types of fields.
To learn more about these skills and exactly what type of fields we are after, please read the mapping rules we have provided.
This picture shows active fields (green outline) and abandoned fields (red outline). Note the recently ploughed look of the active fields, and the vegetation cover on the abandoned fields.
Even if the fields do not currently active or in use, map the ones that may have been in use in recent years. Clearly abandoned fields will look fallow and unkept.
This image shows orchards, which we don’t want to map (red arrows), next to crop fields, which we do want to map (green arrows).
These are plantation forests, which we don’t want to map. The bare patch in the middle with tracks in it looks a bit like a crop field, but it is actually a recently felled compartment of trees, so we don’t want to map it since it will be replanted with young trees.
Here is a settlement that has a mix of crop fields around people’s houses and yards. The fields, indicated by the green outlines, are areas that we do want to map. Although they are close to the homes, there are visible crops growing in the fields. The yards, indicated by the red outlines, are areas that we don’t want to map. These areas appear bare with no crops. The orange outlines show three fields that are hard to tell whether they have crops are not (i.e. they’re ambiguous). We will map every ambiguous field we come across.
Here is another example of a settlement that has a mix of crop fields around people’s houses and yards. The crop fields, indicated by the green outlines, include larger fields toward the outskirts of the houses, as well as smaller fields close to the houses. We will map these fields. The yards are indicated by the red outlines, and will not be mapped. The orange outlines show two fields close to houses that are hard to tell whether they are bare yards or crop fields (ambiguous). We will map every ambiguous case. - The two fields with green arrows show how they will look like when mapped with our interface.
Commercial Farming(Sugrar Cane)
Small-Scale/
Subsistence Farming
Here’s a good example of commercial crop fields (irrigated sugarcane, in this case) right next to small scale, communal farms, where agriculture is undertaken at more of a subsistence level. You can see that the small-scale farmers have left some trees in their fields, and you can even see the really small plots right next to the sugarcane.
1
2
3
4
Lyndon Estes
Associate Research Scholar, Woodrow Wilson School
lestes@princeton.edu
lyndonestes.princeton.edu
Dennis McRitchie
Research Computing Senior Software and Programming Analyst,
Academic Services, Office of Information Technology.
dmcr@princeton.edu
Kelly Caylor
Associate Professor of Civil and Environmental Engineering.
William Guthe
Research Computing Geographic Information Systems Analyst,
Academic Services, Office of Information Technology.
Lecturer in Public Affairs, Woodrow Wilson School.
Drew Gower
Ecohydrology Research Group
Department of Civil & Environmental Engineering
Princeton University
B
N
In order for our project to be a success, we need you.
Although there are computer algorithms to map these fields, they aren’t as good
as the human eye.
To help us, please visit
Once your account
has been registered by Amazon,
You are ready to start mapping
The process works like this:
search for HITs containing the words "Mapping Africa"
Choose to view and accept HITs from the requester “Mapping Africa”.
Identify crop fields falling totally or partially
within the white box in the center of the satellite image.
Fields usually have a semi-regular shape that is distinct from their surroundings,
Fields often have traces of parallel lines (because of ploughing) inside their boundaries.
If there are any, mark each field by drawing a polygon along its boundaries.
If the field falls partly outside of the box, please also map the part that is outside.
Repeat this procedure for all fields within the box.
Submit the HIT. Start another one
We will pay you for each HIT, assuming that your quality score receives at least 60 out of 100 possible points.
How do we determine your quality score ?
Well, every once in a while we sneak in a HIT we have already mapped, which allows us to measure how well your map aligns with ours.
We assess your work's quality
according to two components:
our understanding of where people
are farming is limited.
We have a pretty good idea of farmland distribution in Europe and North America, but we have a lot to learn about other parts of the world.
Of particular note is Africa, which is predicted to experience an explosion in agriculture in the coming decades.
The best data that we have so far are not incredibly accurate.
They are prone to overestimating and underestimating farmland in various locations.
Due to the wide range of error and the unreliability of the data, it can be difficult to understand key issues such as food security, or to predict where agricultural expansion will happen.
As a result, we have launched this mapping initiative to get a better idea of where and how much farmland there is in Africa.
So the process and scoring is fairly simple. However, interpreting what is and isn’t a field can take some practice, and we are also only interested in certain types of fields.
To learn more about these skills and exactly what type of fields we are after, please read the mapping rules we have provided.