Tools and Methodologies-
An evaluation and reflection on Mapswipe and Humanitarian Open Streetmapping.
My experience with the Mapswipe app itself was reasonably solid.
It’s straightforward to use, the tutorial is straightforward, and my initial fears about mislabelling something were assuaged by the fact that multiple viewers evaluate the same piece long before any action is taken with the map information on the ground. In my inexperience, I was wary of missing a house, and perhaps someone might not get treated for malaria because of that. This obviously isn’t the case, but that got me thinking about the disadvantages of the individual model of work vs the crowdsourced, and in the latter how mistakes can be highlighted when there are many eyes on a particular task. I had a few app crashes, but on the whole it worked adequately.
I used the online mode, my limited data means I have to be careful about any image downloading, but the offline mode sounds like a good idea. The activity is gamified via a levelling system to maintain interest over time. I don’t particularly engage with these systems, but if the majority of the user base do, it’s a clever inclusion to increase a user’s time spent on mapping.
On the negative side, there were issues with imagery not loading on occasion, and I feel there may be some detail missed by having the images on smartphones. There screen is necessarily small, making some details very difficult to see, even with the zoom function active. In the Botswana mapping session, I completed multiple passes. Most of the imagery was scrubland with round trees or shrubs dotting the landscape. It became difficult to separate dwellings (also round, and possibly made from the same trees and shrubs) on my small screen. The only giveaway humans were present was from nearby walking tracks, but it was all very imprecise. Such limitations are to be expected and are probably a valid trade-off for numbers of users (more people have phones than tablets, and use them more often), or perhaps the resolution of their imagery is not high enough (perhaps storage or bandwidth limitations, combined with the satellite image itself). Either way, just from a reflection perspective, I felt it was worth noting for potential improvement, even if it is an accepted limitation.
Humanitarian Open Streetmapping
Open Streetmapping is a more sophisticated tool, and I struggled with the detail. With practice, I have no doubt I could master it, but it has a far steeper learning curve than MapSwipe does. It has a focus on detail, down to delineating streets, rivers, and naming buildings. Unfortunately for anyone who hasn’t used software like it, there are quite a few ways to get lost.
For comparison’s sake, I again pursued a Malaria eradication programme. I was required to map buildings in Zambia. Again, a comparatively stable country, so I could set aside my ethical qualms for the most part. In the process itself, most of the unmapped squares, including the first three that I was presented with after random selection, were blank, without buildings. Continuing, I manually selected one that was near another square with a settlement, and drew in what I perceived to be buildings. But again, the lack of resolution at close range really hindered my ability to discern what was a building, and what wasn’t. After saving a few attempts at marking buildings, reviewing them, I ended up reloading the square, and deleting them. Evaluating them on the main map made my ‘buildings’ look like something else entirely, though this was possibly due to a change in light. I was reminded of the ‘Face on Mars’, and realised I was out of my depth. I felt that if I was marking something incorrectly, I was wasting time and resources somewhere along the chain of this project.
Browsing the shortcuts and help sections showed me the level of detail available, but also that I lacked the expertise to use it. (Nodes, wireframe mode, linux-specific controls, radial menus) were all confusing terms to me, and while that points to the more complex nature of the software, I found it to be a barrier. There is far more depth to OHS, but at a cost of turning off people without the requisite experience with similar programmes. This isn’t a software for idly passing time on a train like MapSwipe, but more of a commitment. I think there is room for both, and both benefit in different ways, but OHS was a little beyond me with my current level of skill.
Furthering this, due to it being and open social media system (at least by default), I received a message from a stranger correcting my efforts, but I found it to be quite disheartening. I spent the rest of the time looking around at what other people have done, attempting to validate road systems in Nigeria, but again, felt wholly inadequate to verify anything. I dislike learning in the ‘real’ in areas I’m unqualified for, and particularly ones where people’s health might be at risk from a mistake I make.
While I didn’t produce much in OHS, I did learn something in that area as regards experimenting with real world maps while under-qualified, and even privacy. This mapping activity is now tied to my email. As such, I left editing and validating, and just looked at what other people did until I could understand it better. What I learned from this, is that perhaps this is downside of crowdsourcing, that perhaps I got ‘too far’ into the work without really understanding what I was doing. Did I just validate an incorrect square because I didn’t know how to exit the tab? Would I even know if I did? Could that action harm someone?
These apps tackle the problem of unmapped at-risk areas in Nigeria, Botswana, and Madagascar, (the current top three at time of writing). However, with my initial enthusiasm and appreciation for the app, I paused at the sentence: ‘this information will be available to anyone.’ The campaign to map the Niger province of northern Nigeria was one of the options to map. I know that northern Nigeria is a deeply troubled place, with mass killings perpetrated by Boko Haram in the northeast. There ethnic, tribal and religious fault-lines in Nigeria and serious violence, killings, and mass displacement. If the map information was available to anyone, could I inadvertently be helping violent groups? How would the people in these areas feel about being put on a public map? Has anyone asked them? Do they even know? 2.5 million people have been displaced in the Boko Haram uprising, and many, I would feel wouldn’t want any maps of their area available to the gunmen hunting them. It is an indirect byproduct of this project, but it is a byproduct nonetheless. I have a limited knowledge of North Nigeria, but ‘getting involved’ in areas surrounding a conflict area I know very little about is an issue that certainly bothers me.
While I certainly wouldn’t dismiss these programs, their goal is a noble and innovative one, I would make sure the user is aware that this is a real world activity, and any implications, positive or negative, are also real. A map is a neutral tool, and people need to understand that when mapping a potential war zone halfway across the world.
Learning and my own work
I learned that there is real potential in mixing crowdsourced, amateur humanitarianism with connected apps, and these particular apps benefit agencies like Medicines Sans Frontiere. The idea of mixing new tech and humanitarian work via crowdsourcing if a fascinating one. The benefits are primarily a reduction of cost in labour and man-hours, money which can be better spent elsewhere within the organisation’s work, and covering a vast amount of data much more quickly than individuals could. The development of Mapswipe and crowdsourced efforts is particularly valuable to NGOs and non-profits, as they work in areas where there is limited funding, yet have challenging goals. This approach to mapping bypasses the labour market, and functions on volunteered time. In this manner, my contribution doesn’t require me to work 40 hours a week, it doesn’t require MSF to pay me, and there is a multitude of people from all over the world to contribute when I don’t. There is a great efficiency to this app, to solve a problem via crowdsourcing.
I realised there are ethical questions in getting involved in projects like this. While it’s a noble aim to map these areas for disease and disaster relief, but what concerns me is when the tech proliferates to people with violent agendas. The digital age is changing the face of warfare itself, and I believe it is naive to think belligerents won’t use every available tool to pursue their targets, open community maps included. I’m not sure any real blame can be attached to an individual mapper, (the decision to kidnap or murder is that of the belligerent’s alone), but it crosses my mind that this could be making their operations that little bit easier, and I would have to have more knowledge on the stability of area before my ethical side would be satisfied.
The malaria campaign in Botswana, and my initial choice of Zambia in OHSM were more straightforward decisions for me due to them being more stable countries, albeit of course, in less need. In geographical area choice of project, I feel there are ethical questions. The unfortunate truth seems to be that the areas with most need are frequently the areas where people are fleeing persecution, and at the most risk if they are found. “Could I potentially do harm here?” is something I asked myself, and I couldn’t puzzle it out, despite the tiny input I might have in the operation as a whole.
Another area I gave consideration to after my experience was dynamic of the amateur/crowdsourced method versus the professional/individual method. The crowdsourced model takes in vast amounts of data, but I feel sometimes is vulnerable to amateur mistakes, and perhaps outright malice, on occasion. While systems are in place to verify the information, time is still wasted on this, and, as I mentioned, if the verifier is also an amateur, or malicious, then problems could occur.
This exercise has made me more aware of the possibilities of both methods, and while I can’t really apply crowdsourcing to my project, I liked learning about the possibilities. My work is in video, and while I have been using digital maps to scout locations for a while. On a very low level, putting out requests for locations, props, information over Facebook and twitter might count as crowdsourcing.
I browsed my own area, while is fully mapped on Open mapping, and using it for location scouting would be my primary use for it. There might be a benefit in a group marking more rural scenic places that are not present on the main map due to remoteness, a small crowdsourced effort could generate something worthwhile there, and I could see this as a potential use for me in future.