Maps 2.0

Communications desk. Incoming reports were posted on Google Earth/Google Maps by remote volunteers, then displayed in the emergency coordination center (ECC).

On April 27, a virtual team of volunteers from HumaniNet and GISCorps, working remotely, entered reports generated by Multnomah County hospitals and transportation teams in Portland, Oregon, during Exercise Cascadia Peril.

The process was straightforward: reports arrived in the emergency coordination center (ECC) by radio (see photo) and were submitted to the on-site HumaniNet team for entry into Trac, our collaborative workspace.   They were then picked up by five “mappers” in four states, converted to kml files, and returned to the on-site team for posting in a master map created with Google Earth (GE).  This was displayed on ECC monitors, side by side with the maps created by the county GIS team.  The map on our exercise page is the Google Maps (GM) version.

Below, our Portland team (Matt Blair, Sue Gemmell, Marcelle Caturia, and I) will comment on the lessons learned during and after the exercise.

Our thanks to Dave Houghton, director of county emergency management, and his  team for their cooperation and the  opportunity to participate.   Dave believes that this is the first time that a remote volunteer mapping team has supported a disaster response exercise in the U.S.   We believe that the concept works, and we invite you to review our findings and post your own questions and comments.

In a separate post, we will ask the five terrific volunteer mappers to write their comments.  They are: Michele Cote and Joe Ludwig of HumaniNet and Mary Meade, Mike Price, and Paul Trudt of GISCorps.   Sincere thanks also to Shoreh Elhami of GISCorps for her outstanding collaboration!

Initial observations

From Matt Blair:

1. An on-site coordinator could send data reports to a remote support team outside the disaster zone, and they can return useful situational awareness maps over a minimal, satellite-based internet connection.
2. Using off-the-shelf, low-cost and free tools, a remote team of volunteers can identify and label critical infrastructure in a city they don’t know well, if at all.
3. It is critical to give careful thought to what kind of information you need to collect in each status report, and how that information can be depicted on a map, before the disaster starts.
4. The reporting date and time, reporter’s name and contact information should be required for every report submitted. This allows decision makers to understand how current the information is, whom they should contact for an update, and how.

From Sue Gemmell:

1. Team must be able to stay focused on the ultimate purpose of the maps.To stay focused on the purpose, it is necessary to know who will use the maps, and the decisions and actions that will be supported by the maps. This focus will ensure that the right data model and process can be supported as it is modified during the response. My recommendation: at the initial outset of the response, write user profiles (a paragraph or two). For each profile, identify actions and decisions (scenarios) that the maps will support. For those actions and decisions, identify the data that is required. As the response unfolds, changes to data model and process will be needed. The mapping team will be able to use the profiles and scenarios to decide whether and how to make changes by answering the question, “will the changes to the process/model support the end user’s ability to make decisions and perform actions?”
2. The information gathered in the reports needs to be unambiguous.
   The report forms (especially the hospital form) need to be analyzed and redesigned so that the content supports the ultimate purpose of the maps, as described in comment #1. My recommendation: review reporting tools in conjunction with user profiles, actions and decisions, and data needs. The reporting tools must align with the data needs.

In conclusion: a little planning up front – review of purpose and tools – at the onset of the response – will ensure the best value of the response.

From Marcelle Caturia:

Flexibility was key in matching volunteer roles with particular tasks.  For example, initially we had two volunteers filling the role of “Task Coordinator”, which required both: 1) entering data from field reports, and 2) assigning tasks to remote mappers.  Soon it became clear that dividing tasks 1 and 2 between the two volunteers would streamline the tasks immensely.

Portland (part of Multnomah County). Click for live map.

On April 27, HumaniNet’s Maps 2.0 team will support a major disaster response exercise in Multnomah County (Portland), Oregon, by posting essential exercise information on a Web-based online map.  HumaniNet will also provide assistance with satellite communications in the multiagency exercise, augmenting landline and radio communications among participants in northwest Oregon.

The mapping and communications events are the first in a series of simulations in the coming months under the exercise name Cascadia Peril, which will begin on April 24 with a scenario of a magnitude 9.0 earthquake followed by a coastal tsunami. In subsequent phases, exercise activity will include several counties in northwest Oregon.

The initial phase, which is being planned and led by the Multnomah County Office of Emergency Management (MCEM), will focus on hospital reporting to the emergency coordination center (ECC) and mapping of events and field reports of affected transportation infrastructure, such as damaged bridges and overpasses. Following a multi-agency amateur radio drill on April 24, county agencies, including Multnomah County GIS and Transportation, will participate in the activities of April 27.

The HumaniNet Maps 2.0 team will work closely with MCEM in posting time-critical information on Google Earth and Google Maps. The online map will be accessible beginning on the morning of April 24 at the HumaniNet home page, www.humaninet.org, enabling participants and all interested citizens to observe the exercise activity on a dynamic map. Team observations and photographs will be posted on the Maps 2.0 blog.

The HumaniNet team will work closely with MCEM and the county GIS team, who will manage information using specialized mapping tools.

The goals of the mapping activity in late April are to test the parallel processes for tracking disaster data using both the county GIS tools and the HumaniNet – Google Maps tools. The Google Earth Outreach team, an initiative of Google, Inc., will continue to provide valuable technical support.

The Maps 2.0 team is formed primarily of volunteers who are trained in map-based data entry and who are committed to support Oregon officials throughout the summer exercise. In addition to the HumaniNet volunteer team, GISCorps is seeking volunteers to assist in the exercise and to take advantage of the exercise to observe and research information processes in an emergency.

The Maps 2.0 planning for the Cascadia Peril exercise builds on the experience of the HumaniNet team in support of Exercise Talom in Thailand in November 2008 (see blog post below).

November was a month to remember for the Maps 2.0 team.  Our research into Google Earth and Google Maps (GE/GM) convinced us that these tools would make possible a user-friendly, reliable way to display critical information in a disaster response or other emergency.  While others have used GE/GM in emergencies (e.g. the California wildfires in 2007) to  our knowledge no one has attempted a planned operational process for posting information on online maps in support of humanitarian teams, and with their collaboration.  (If we are wrong, write a comment and tell us about it).

In the picture below, Kyaw Oliver  and Hector Carpintero, both of ADRA, are online over a BGAN satellite terminal at the team operations center in the village of Ban Yang Loung in northern Thailand.  The GPS coordinates that the ADRA and World Vision teams collected were posted by the HumaniNet team within hours of receipt, and the teams are able to view the results over satellite.  You can see the same maps on our Thailand exercise page.

While we learned a great deal about managing information and collaborating with a “virtual team” (including volunteers) on the other side of the world, we all agree that the Google tools work, and it is very possible for non-tech people (like me) to learn how to use them.

You can read more about Maps 2.0 on this site and the HumaniNet Web site, where you will find some background and links to the Nepal simulation in late 2007.

Leave a comment, tell us your success stories, ask questions!  While we are not all GIS professionals, we’re excited about the potential of displaying important information “geospatially,” in a way that your NGO or nonprofit team and supporters will easily understand.  People want to know where you are, where your teams are operating.  It’s not hard to do – build it into your 2009 plan.

Gregg Swanson

We did a webinar on Google mapping tools last week, with participants from Mercy Corps, HISG and Engineers Without Borders.  The slides are available on this page, and we’ll be posting video online soon.

by Matt Blair

The emergence of free and (relatively) easy to use mapping tools has created an explosion of mapping activity, and the results vary widely in terms of usability.  

As part of our research, we have been trying to identify some of the attributes that make maps effective.

Here is a preview of some of our suggested practices:

• Decide what questions you want the map to answer before you start.  Review frequently as you collect data and build your map.
• Answer one question per map.  If you are generating KML, you can always combine data files and features later.
• Coordinates are just metadata.  Don’t forget the data! Be sure to address questions like: What happened? Is this different from the norm? Why is it important?  What should we do about it?

In addition to coordinates, critical metadata for a placemark includes:

• sensitivity and security
• source
• accuracy
• description
• reported and “valid until” timestamps
• contact information and/or links to original or additional data

Crowd-sourced data is often incomplete. Simply adding these last two metadata elements could make a huge difference. Any connection of data back to the real world can make inert data actionable.

To expand on that last point, a placemark on a map after a hurricane that is labelled “Shelter”, with a balloon that says “Capacity: 200 beds” is of limited use.  Do they still have 200 beds?  How many people are there now?  Are they up and running? Do they have power, water and supplies? Have they shut down already?

When there just isn’t time for detailed data collection, the most critical information to include is the contact information:  who can I call or text or email to find out the current status?

Stay tuned for additional suggestions on how to make your maps effective.

Update (10/14): I’ve expanded on these ideas in section five of our recent webinar.  Slide are available here (PDF, 15.8 mb) with video to follow.

by Matt Blair and Gregg Swanson

If the recent conflict in the Republic of Georgia sent you scrambling for some perspective on Google Maps, you may have been in for a surprise: no roads or railroads, no cities, no features that would help make sense out of a dynamic situation in a country of over 5 million people. Even Google Earth, which is typically more comprehensive, had only cities and a few wikipedia links, and some postings of tourist photos, but no additional infrastructure.

According to Google, they were not able to find data that met their quality standards for Georgia, Armenia or Azerbaijan.

Is a map of questionable accuracy better than no map at all? That’s one of many questions we are exploring as part of our Maps 2.0 initiative.

Nearly two weeks after the conflict began, Google’s view of Georgia is still empty. In the meantime, ReliefWeb is aggregating static maps of Georgia from multiple sources. As an experiment, Matt overlaid one of the ReliefWeb status maps  on a Google Earth (GE) map of the region. (Click here to view the overlay in Google Earth.)

While the base information in GE is minimal, the value of overlaying critical information on a good quality map (even one that shows only satellite imagery) is clear and compelling. In this case, information on movements of displaced persons could be added, updated, and managed as the user selects the layers that best serve their purpose.  

In our recent interviews with humanitarian managers in the field, we are finding that easy and fast access to current regional and local information is an acute challenge that humanitarian teams have faced for years. The Google Earth and Google Maps toolkit and similar map-based, Web-delivered applications have nearly solved the accessibility problem. The next step is the “fast” part – to design and implement a process for loading critical information quickly and efficiently on those maps.

Stay tuned – we will have more updates on this research in September.

Maps 2.0 has been made possible through a generous grant of the Meyer Memorial Trust and contributions from our other donors. You may help support this project with an online donation at our “Donate Now” button.

Last week, the ADRA-Caritas team in charge of the upcoming Nepal simulation, Exercise Khichadi, decided to postpone this training event, since the worst monsoon flooding in many years (perhaps ever) was clearly going to occupy relief teams and managers for some time to come. The simulation will take place in early November, preceded by the rescheduled Tech Day on October 31.

The rescheduling activity has compounded the difficulties that the many relief organizations in Asia, including ADRA and Caritas, are encountering.

In parallel, a few volunteers with strong expertise in GIS are working with our partners at ESRI on a small-scale GIS demo for the simulation. Volunteer Sue Gemmell (Portland, Oregon, USA) is coordinating this project, working closely with David Gadsden and Scott Moore of ESRI, in Olympia, Washington, USA. Please watch our Nepal GIS page, and those of you with an interest in humanitarian GIS may wish to RSS this blog, so you will be notified of new postings.

Since the NetSquared conference in May, we have expanded the community of interest and discovered a number of initiatives that are making real progress toward GIS support for situational awareness (SA) in field operations. We will highlight the “best of class” in this blog over the coming months.

Since our team has not tested any of the solutions, we cannot yet say which are ready for field use. But we do see a debate taking shape along numerous axes, including:

• Open Source (FOSS) vs. commercial software
• The problem of data entry
• Simple (pure location, e.g. where is the team located) vs. sophisticated (analytical tools, access to multiple data layers)
• Focus on GIS, or information management? Or both?
• Involvement of program users, not just the ICT experts
• Extent of real-world testing by real-world field teams

We will invite our volunteers and guest contributors to weigh in on these and other topics, with the intention to spur a vigorous dialogue and shed light on GIS solutions that work. This means solutions that work in a field environment, without costly and not-so-mobile infrastructure and designed for quick training, ease of use, and local adaptability.

The demo being prepared for Exercise Khichadi will be simple and modest. Our purpose is to stimulate discussion and collaboration where possible.

My first test is one that I heard from a team in Afghanistan after they returned in late 2001: “we could not find where the UN meeting was to be held, or what time.” Understandable, in a mid-sized city in an unfamiliar culture. Other possibilities and necessities (security, medical, logistical, etc.) were beyond the reach of this team, even without connectivity and dynamic maps, if they couldn’t find the people doing the coordinating. Much has been done since 2001 to remedy this, we are told, and that is what our team will be looking for.

As always, comments welcome!

Gregg

By Gregg Swanson, Executive Director, HumaniNet 

It’s been a full week since returning from N2Y2, and I’m still processing my notes from dozens of conversations and some valuable (and somewhat intense) sessions at N2Y2.   Our calendar this last week has been jammed – a (big) birthday, a (very big) graduation, and some exciting developments in the core activity of HumaniNet – to be covered in a subsequent post. 

Matt Blair and I met some outstanding people from the NetSquared and nonprofit communities, and we send our sincere thanks to the Net2/TechSoup teams for planning and hosting a memorable, enjoyable and high-energy event.  It was a learning experience, for sure. Although Maps 2.0 was not one of the top three winners, we were truly honored to be one of the 21 Featured Projects.  It was fantastic to meet and learn from a group of very sharp, committed nonprofit leaders and funders.   Their organizations are certainly in good hands.  We would not hesitate to collaborate with any and all of them. 

Our discussions abundantly reaffirmed the need for a GIS resource for nonprofit and humanitarian managers, planners, and researchers.   Here is our bottom line: we will continue to grow the GIS community of practice, but in a paced, methodical way.    

The Maps 2.0 team is sorting through the many possible approaches and will launch a project that can and will be sustained.  This is the HumaniNet model, as it unfolded since 2001 – one building block at a time, taking care of field partner needs and creating user-driven processes and content. 

Here are a few summary comments and observations from N2Y2, in random order: 

*  Leaders of technology-based nonprofits fully understand the power of networks, and social networking.   This includes Net2, TechSoup, and Compumentor, but also visionaries like our friends and partners Allen Gunn of Aspiration, Kim Lowery and Tobias Eigen of Kabissa, and Randy Roberson of H.E.L.P.  Actually, all of the 21 Featured Projects and their leaders – and probably most of the 152 original entries – are built on cooperative and networked strategies.

*  When you are on the playing field with star projects that deliver direct assistance in such areas as recycling, family support, and civic engagement, it is not that easy to explain to funders how an enabling model (such as HumaniNet’s) will effect social change.   That’s one of our primary challenges with Maps 2.0.  

*  It is still a bit surprising, but true just the same, that some folks who pay close attention to the nonprofit world are puzzled by the idea of “shedding light” or providing “consumer guide” content to humanitarian teams.  In business, consultative assistance and current tech information is taken for granted.  Excellent resources such as TechSoup, Idealware, and Aspiration (see their Social Source Commons) work really well for nonprofits.  Why is it different if you are managing a refugee camp in Darfur or a medical project in Sri Lanka?

*  A few years ago, I met an executive who said “humanitarian organizations will never collaborate.”  He was dead wrong.  Since the Maps 2.0 vision and initial plan was formed, just two months ago, we have heard from many NGO managers who are eager to share information in this booming space.   There is a sense that GIS will help “raise all the boats,” which is the idea with collaboration.

*  Another challenge for the Maps 2.0 team is to define the primary use cases (“verticals”) and to determine which should be addressed initially.   We will address this by paying close attention to the different user communities.  And not just GIS superusers!  An NGO manager in Nairobi who has had no exposure to GIS still has a set of information needs in running her or his project, and we will reach out to people who are “heads down” in the field.

*  The importance of mapping tools and data sets is massively clear to many NGO program and tech managers.   We’re fortunate to have the advice of several of them on our Advisory Team (which see).  We have been contacted by several highly qualified GIS experts who have offered their assistance in research and evaluation.

*  We also met some very forward-thinking managers from Silicon Valley tech companies, who understood at once what we are about, and the power of networking.  It is really encouraging to know that corporate citizenship is at work in the tech world, big companies and small, software and hardware, local or multinational.

Please check back soon – probably by June 21 – for more on the Maps 2.0 planning and an announcement of a proof-of-concept field demo for relief operations, tentatively scheduled for late August.

Once again, we thank the Net2 team for a fascinating and network-intensive two days!  Also to Gunner for his leadership of the pre-conference meeting, and to all of the Featured Project leaders for being accessible, helpful, and collaborative!

Gregg