Crowdsourced Spatial Project Participation

Crowdsourced spatial projects are projects were users collaborate in the creation of spatial content that is made publicly available. This article looks at two of these projects, MapSwipe and Humanitarian OpenStreetMap Team (HOT), and reflects on the experience of taking part in them.


MapSwipe is an app of the Missing Map Project that let you go through satellite imagery and mark tiles that contain relevant features (building, roads, etc.). The imageries are from areas that have some humanitarian crisis and need to be mapped.
The app is easy to use; there is tutorial that explains how to identify the features and a reminder of how to perform the task at the start of every session; it explains how to mark the tiles, how to zoom a tile and how to move to the next set of tiles.
When I tried the app there were three missions open (Fig. 1). I participated in each of them.

Missions in MapSwipe
Fig. 1 – Missions in MapSwipe

The quality of the imagery varies even within the same mission; it is usually fine, even if sometimes there are few screens covered by clouds. The task is quite easy: just tap on a tile if it contains one of the relevant features making the tile green. The only difficulty is identifying the relevant features as sometimes they are not so evident, especially the round buildings; but there is the option to tap twice when not sure of the presence of the feature (the tile becomes yellow) (Fig. 2); that takes out the anxiety of missing a feature or reporting a feature that is not actually there.

A 'maybe' tile in yellow in MapSwipe
Fig. 2 – A ‘maybe’ tile in yellow in MapSwipe

The app is a good idea to get people engaged in a sort of playful activity, while contributing to a humanitarian cause. There are levels to reach to make the app more game-like and keep people engaged and, while you are doing the task, the app shows encouraging messages to keep you motivated (even if I found the messages a bit annoying).
There is the option to download the map and work offline; this is a good way to have people working at the tasks in their spare time without having to worry of spending money when they are not connected to a wi-fi; however I did not try this option.

Humanitarian OpenStreetMap Team (HOT)

The Humanitarian OpenStreetMap Team (HOT) is a project where users participate in drawing maps for places with humanitarian needs in order to facilitate the work of humanitarian organisations.
The HOT website manages the tasks for the areas that need to be mapped; while the drawing of the maps is done on the Open Street Map (OSM) map editor. So you need to be registered with OSM to collaborate in the project. It is also advisable to do first some mapping in OSM in your own neighbourhood to get some practise with the editor before starting to engage with the HOT tasks. I did that and did all the tutorials on MapGive. I then went into the HOT Tasking Manager to choose a project. There were 130 pages of projects; and that gives an idea of the amount of work that needs to be done. Each project has a priority and by default they are ordered in order of priority; so you get to see first the project that are more urgent at the moment. I chose the “#2563 – Missing Maps: Zambia Malaria Elimination 12” as it was one with the highest priority at the moment. It is one of the mapping tasks in support the Clinton Health Access Initiative’s malaria program that aims to put in place interventions to reduce children mortality caused by malaria. I read the instructions for this specific tasks and moved to the selection of the tiles to map. I noticed that only some areas needed to be mapped (Fig. 3); I supposed that these were areas where features had been identified with MapSwipe; but I could not find any confirmation of that.

Mapping areas for #2563 - Missing Maps: Zambia Malaria Elimination 12
Fig. 3 – Mapping areas for #2563 – Missing Maps: Zambia Malaria Elimination 12e

The mapping task was for buildings only and the mapping process was quite fast as most of tiles that I selected had no buildings. While some building where easily identifiable, others where more confusing; especially the round building that could be confused with trees. There is no “maybe” option as in MapSwipe, so I was left with the doubt if it was better to map less and miss some buildings or map more and report as buildings features that were not actually buildings. I would suppose that mapping more is better as at least would not completely ignore people that may be in need of help; even if this could lead to a waste of resources. Anyway most of the times I just choose to not mark the tile as complete when I was not sure. The doubts were even more relevant in the validation process; especially as only one validation was enough to mark the tile as validated and displayed in green; I would have expected that the validation process would be done by more expert users and that a certain number of validation would have been needed to mark a tile as validated; but that was not the case.

Mapping buildings
Fig. 4 – Mapping buildings

Another thing that can cause issues in the mapping process is the fact that the logging of the activity on the task, that is made in the HOT website, is not synchronised with the activity in the map editor on the OSM website; so you can unlock a tile, and even map it as complete, without having saved the work on the map editor. In the validation process I found a tile, marked as complete, that was clearly missing a lot of the buildings; I found it quite strange that a mapper could have missed so many features, but that could have been due to this lack of synchronization and the mapper had forgot to save before marking the tile as complete.
This task uses imagery from DigitalGlobe’s Maps API Premium Imagery. In the OSM map editor it is possible to change the background in use; I changed it to the Bing satellite imagery (the default in OSM) to compare the two different satellite images. Apart from alignment issues, I noticed that in some cases the buildings were quite different, so I could get an idea of how the dwelling area had changed with time (I suppose that the DigitalGlobe’s imaginary is more recent, but could not find any information about that) (Fig. 5 and 6).

A group of buildings in the DigitalGlobe's Maps API Premium Imagery
Fig. 5 – A group of buildings in the DigitalGlobe’s Maps API Premium Imagery
A group of buildings in the Bing satellite images
Fig. 6 – A group of buildings in the Bing satellite images


I am so used to rely on Google maps or other type of maps, that I never thought that there could be parts of the world that were actually not mapped at all. I went to look at the area that I was mapping on Google Map and it was disheartening to see the area all white with only the road traced (Fig. 7).

Zambia area on Google Map
Fig. 7 – Zambia area on Google Map

These gives an idea how even something like a map is connected to money; the area that Google maps are areas that have some business interest (shops, restaurants, companies, etc.) and so are source of potential revenue; the other areas are left out; if you looked at them you would assume that there was nothing; while there is community activity, only that community is of no business value. This highlights the importance of the HOT mapping projects to put on the map these communities that have “only” humanitarian value.
On the other hand one thing that has worried me doing the maps is: what if these people do not want to be on a map? On the HOT website there is a profile of the project with information about the organization that needs the map and their purpose; but then the maps are publicly available on OSM and everybody can use them. What if they are used by totalitarian governments for repressive purposes or terrorist groups to carry on attacks? Of course it is always a matter of considering the advantages and the positive improvement that the drawing of the maps are facilitating against the potential drawbacks of the usage for negative purposes; and sometimes you have just to accept this possible risks.
At last a consideration about the potential information can be learned about places by looking at the satellite imagery. When we travel to a place, we learn something about the habits, culture economy of the people living there. Now we can travel all around the world like on an airplane just by looking at the satellite images and infer information about the places. For example in the satellite imagery of Zambia that I was mapping there were some circles (Fig. 8); they are central pivot irrigation circles (I had never heard about them) and from them it is possible to deduce information about the type of agriculture that is practised in the area.

Central pivot irrigation circle in a satellite image of Zambia
Fig. 8 – Central pivot irrigation circle in a satellite image of Zambia

Leave a Reply

Your email address will not be published. Required fields are marked *