OpenStreetMaps Mapping Party, Dunbar – 23/03/13 – A Review (Part One)
On Saturday the 23rd of March, I made the cross-country trip to the beautiful coastal town of Dunbar to take part in the first OpenStreetMap Scotland mapping party of 2013. I was invited to make the trip by Tim Foster, regular OpenStreetMap contributor and fellow member of the emerging Stalled Spaces project with Bruce Newlands at Maklab.
As I have previously discussed, mapping parties are a method that the Stalled Spaces project might explore as tool to collect data and foster a community of interested individuals. I was keen to get to grips with what was actually involved in the activity and was lucky enough that one was scheduled so soon after the initial Stalled Spaces meeting. With my researcher hat on, I braved the late-winter cold snap we’ve been under to determine what the Maklab project might take from mapping parties, as well as their wider implications for peer-produced and open source projects like those I am researching for my MPhil.
Meeting at Dunbar’s new Bleachingfield Community Centre at 10am, the day began with introductions and a setting out of a rough itinerary of activities from senior OpenStreetMaps contributor, Bob Kerr. I had a small amount of knowledge of how OpenStreetMaps was edited and maintained, as well as some of the advantages and challenges that the system has compared to traditional commercial mapping organisations. I had taken a look at the already fairly highly detailed OSM map of Dunbar in advance. However, it was enlightening to hear Bob explain the aims of the day, as well as give some well-informed background to those in the room who had little previous experience of OpenStreetMaps. One particularly illuminating description was of the Mapcake that had been established in advance of the day. This tool allows the map to be partitioned so that contributors can select an area they wish to work on. These segments are colour coded to show their relative state of advancement. A number of people had spent several days before the mapping party, updating the map from aerial and satellite photography, so that there was a sufficient base of data from which to build the day’s activities around.
Next, Tim gave those who were interested an introduction to Potlatch, the browser-based system that OpenStreetMaps uses to edit the map data. As someone who had very little experience with the system before, it was a fascinating insight into a software that seems highly intuitive and user-friendly. One aspect of OpenStreetMaps success seems to be how well it engages those who are not from highly technical backgrounds, allowing practically anyone with an internet connection to add their contribution. While the depth and complexity of the map seems nearly infinite, there exists an ability for more casual users to drop-in and add positive and consistent contributions (minimising, and even negating, what Brooks would describe as the “ramp up time” associated with new contributors becoming involved with a technical production, where the effort to train new members out-weighs the gains they eventually bring to the project, a part of the phenomena dubbed Brooks’s Law¹).
While that introduction was taking place, a few of the more seasoned mappers divided up the map and set out with GPS devices, cameras and good old-fashioned pen and paper, to gather data on areas of the town. A number of the group had come with bikes, eager to detail the local cycle network and woodlands. I can imagine that there is a strong connection of motivations for those attending mapping parties; the desire to improve the map and equally the desire to experience a physical environment from a new or unusual perspective.
While the gang of intrepid cyclists braved the blustery conditions, Tim took the remaining newbies for a walk around the community centre. As it turned out, one of the areas on the map that was in need of updating was the building we were based in for the day. As a great deal of the mapping is reliant on tracing publicly available aerial photography, it is unsurprising that relatively new developments are often missed until they can be verified by actually exploring the land itself. Without any more detailed reference data to draw from, we took our phones and GPS devices outside and walked around the building so that we could use these traces to update the map. By the time we were finished and had delved into Potlatch, it was time for lunch at Dunbar’s excellent community owned bakery.
After reconvening to establish next steps, more of the group headed out to explore and map. Bob asked if I would be able to help him scout out some of the southern extent of the town where there is a great deal of new housing development which is still under construction. One of the ways in which the OSM map’s relative development is measured against traditional maps like the Ordnance Survery, is by the relative ‘completeness‘ of the names of streets. A register of the number and position of street names listed on OpenStreetMaps is measured against one that is released by Ordnance Survey. At the time of writing, the OpenStreetMap data was 95.29% ‘complete’ (a wittier blogger would probably be able to throw in a joke about Jorge Luis Borges around here…).
Bob and I jumped into his car, with a video camera to record our commentary on the places we passed and drove a fairly comprehensive loop around the town, focusing on under mapped areas. After around an hour, which included a stop at a golf club in search of a course map and a spot of very near off-roading through the very newest ‘streets’ in Dunbar, we returned and Bob sat down to review the recording and update the map.
Thinking of what I could specifically bring to the proceedings from my own background in architectural design, I had a small experiment with ripping data from the OSM map and running it through SketchUp. While my results were fairly rudimentary, Tim and I discussed the possibilities for exporting data from OSM for use in 3D modelling (both digital and physical). This is something that is central to the Maklab Stalled Spaces project. The ability for data to be quickly taken from OSM to fabrication technology seems an exciting new avenue that this project could push. There are a few systems already in use that simply output OSM data as quickly rendered 3D maps (including OSM-3D) which pulls tag data from OpenStreetMap to generate the shapes of buildings. A significant challenge for these maps is the inclusion of topographical information that adds elevation data. There exists no highly-detailed, public domain source for elevation data that can be globally used. While it is possible for such information to be added to the map, due to this lack of widespread source data, it is not a significant feature of OpenStreetMaps. Instead, a number of projects (OSM-3D included) have made us of data released by NASA, taken during their Shuttle Radar Topography Mission. While this is an incredible, almost sci-fi source for public domain data, unfortunately, this data does have significant limitations in its accuracy at the scale of urban settlements, a scale that would be required for the sort of model making we are interested in. It is certainly an interesting start, and there is a debate to be had about how necessary such information is at all for the Maklab project.
As the day wrapped up and a few Edinburgh and Glasgow mappers boarded a train back west, It was interesting to chat to a few now fellow OpenStreetMap contributors. It seems, at least from this one encounter with the OSM community, that there exists a strong relationship between members personal interests and their professional lives. A large proportion of the group had jobs that were in the fields of GIS and mapping generally. Some were regular map contributors and attendees of meet-ups and mapping parties. Others, particularly those from Dunbar and surrounding areas had associations to local community groups, such as Sustaining Dunbar, the group who had brought the OpenStreetMap mapping party to the town for the day. I’m interested in exploring further the identities of those who contribute to OpenStreetMap, their backgrounds, their relationships to the spaces and places they map, and the community they form. Hopefully I’ll return to this subject in a later blog post.
I will definitely write further on the mapping party. I have at least two follow-up posts planned that explore some of the issues I picked up on while taking part, and the implications that I believe they might have for the Maklab project, as well as the development of open-source/peer-production within the field architecture. One of these will focus on the novel techniques that OSM uses for coordinating labour and resources in the development of the map, both at the mapping party and more generally through the online editing platform. Another will explore the mapping party as a novel form of community engagement tool for architects and other fields involved in the development of the urban realm. There was a genuine excitement that could be felt in the room during the mapping party, and there seems much promise in exploring how mapping and other hands-on activities can engage locals in the planning and development of urban areas.
For now, my thanks to Tim for the invite, Bob for a true taste of hands-on mapping, and all those others I met on the day. Morag Haddow of Sustaining Dunbar has set up a Facebook group to keep those interested in contact, and there is now an OpenStreetMap wiki page for Dunbar that includes details of a proposed future mapping party on Sunday the 16th of June. Hopefully, by that stage, the Maklab project will be bearing some fruit and perhaps Dunbar represents a test-bed for some exciting ideas. Certainly, it already has the map to make a great start!
¹ It should be noted that I’m not suggesting that Brooks’s Law is applicable to the development of OpenStreetMaps, as the project is neither software development (which typically involves the management of more complicated data) or one that could be described as ‘late’ a stipulation of Brooks’s Law. Rather, I was impressed by the ease of getting to grips with Potlatch, which seemed simple enough for almost anyone to teach themselves to use. Specifically, the activities of the mapping party (where time was limited) did not seem to be greatly inhibited by new users draining time from more experienced members.