Liberia: lessons we're learning

In the five months since our team began implementing the Ushahidi platform in Liberia, we have been constantly learning.  Whether it’s from our partner organizations, the frequent power outages or our own expectations, we’ve cobbled together some lessons that we realize aren’t worth much unless they’re shared. To provide some context before diving in, here’s a synopsis of our latest activities: -       There are now 15 customized Ushahidi instances up and running in Liberia (go to Liberia Mapped and click on the category “Using Ushahidi” to see who’s doing what), a couple of which are private due to the sensitivity of the information.  As of late November, a total of  over 3,900 reports have been published on these customized instances by our 12 partner organizations and citizens at-large. -       The most recent of these instances is Liberia 2011, a map for presidential election scheduled for October 2011.  More than 10 election monitoring organizations and institutions will be contributing to this map once it is fully launched in late December.  In order to encourage reporting, we established the shortcode 2011 for SMS reports; it is currently free for all customers of Lonestar, the country’s largest telecommunications company, and we hope to have all five operators onboard within the next month. -       Some of our recent tech additions include:

Timeline plug-in – This plug-in allows users to plot events with a start and end-date on the map so not just single incidents are visible on the timeline.  This plug-in also allows administrators to choose if they want days or months displayed in the timeline.

Google Earth Layers plug-in – This plug-in allows the user to choose between Google streets, hybrid, map maker and map maker hybrid layers as the base.

SMS automation: This plug-in intercepts incoming SMS messages, checks for the required keyword and proper formatting and, if they pass, automatically approves and maps them.

Java program for Frontline: this program ensures that messages will be sent from FrontlineSMS to Ushahidi even when the internet is unreliable.

HTML-friendly text box and document upload: The text box used for entering reports is also now HTML-friendly, so specific segments of Word documents can be copied within the text box and will retain the original Word formatting.  Also, it is now possible for partners to upload Word documents or PDFs to specific reports and to the instance’s main page.

Lessons Learned Learning From Our Partners In the course of training partner organizations, our team has learned that many of them are well-connected to communities in nearly all of Liberia’s fifteen counties and are already collecting early warning data in the course of their peacebuilding activities.  While these may sound like ideal prerequisites for Ushahidi users, when introducing partners to the platform we’ve found that several have been slow to utilize the tool.  In fact, seventy-five percent of current partners upload less than two reports to their maps each week and have a critical mass of ten or fewer reports total. Part of what these low reporting rates tell our team is that partners aggregate, analyze and share information differently.  For some organizations, the Ushahidi platform is a natural complement to their current methods, while others are not accustomed to geospatially visualizing information or reporting conflicts via SMS.

An example: the Norwegian Refugee Council and the Carter Center have hundreds of reports on their Ushahidi-Liberia instances; these are also organizations that have been strategically collecting data for months in well organized excel spreadsheets that can easily be uploaded to Ushahidi with only a few adjustments.  When you walk into the NRC and Carter Center offices, there are maps on the wall; these organizations have been visualizing and collecting information in such a way that Ushahidi is, in a sense, just what they’ve been waiting for.  Some of our local partners, on the other hand, have a handful of mostly volunteer staff that either 1. do not have a systematic approach to gathering and storing data, or 2. catalogue information in a way that is not easily transferred to the Ushahidi platform.  Many of our local partners focus primarily on producing an annual report that summarize activities for donors – a document that is typed and sent once, rarely to be referenced again.  There are no maps on the walls of these partners’ offices (more on this in just a moment), and in some cases the concept of aggregating and storing information for analysis and sharing is little more than a funding requirement.  Before asking partners what they want to map, we should be asking for an introduction to their data collection and storage habits. Another factor that accounts for underutilized instances is the technological proficiency of many partners’ staff.  The combination of tools the platform employs is new to most staff we train.  The unfamiliar configuration of technologies wouldn’t be a challenge if partners’ staff were well-versed in each of these tools or, more importantly, were educated to think critically when learning something new. It is clear from our initial trainings that many of our partners’ staff rely on memorizing each step rather than learning how all the steps add up to create a whole.  We’re in the final stages of releasing a manual for partners on Ushahidi in the Liberian context, but we are up against the much larger challenge of Liberia’s lacking education system – one that emphasizes memorization over deductive reasoning.  To put Liberia’s schools in context, consider the following: the West African Examinations Council (WAEC) administers an exam to high schoolers in West Africa that they must pass in order to attend college.  In Liberia, this test is taken by students at UL’s law school (considered the brightest students in Liberia); last year, only one student at the law school passed the WAEC exam, and with a near-failing grade.  Visiting professors from the American Bar Association have begun using games like Sudoku at the law school in order to cultivate critical thinking. It seems that, in order for all partners to develop a familiarity with the Ushahidi platform, our team will likely have to teach critical thinking alongside crowdsourcing. Our team has also learned that partners might use Ushahidi more regularly if they understood the benefits of mapping information.  We assumed that the utility of geolocating events would be self-evident based on our own experiences of using maps in the US for everything from finding nearby restaurants to live maps of presidential election results.  However, in a context where maps do not serve even their most basic purpose of navigation (there are only a handful of main roads in Liberia and one need only ask passersby for directions), maps are not considered inherently useful tools.  This does not mean the Ushahidi platform isn’t useful for partners; instead, it suggests that some partners might develop a new perspective on geospatial information with specific trainings on the many ways maps are used around the world today. In addition to the challenges of learning new tools, many partners’ staff are overtasked with their daily workload.  All of our partners have expressed enthusiasm about Ushahidi, but few partners have designated specific staff to implement the platform and fewer still actually have staff that are available to take on a new responsibility.  Further, nearly all partners are limited by unreliable power and slow Internet connection; for some organizations, it can take more than five minutes to load an Ushahidi instance.  At least two partners have only one USB modem for all staff to share.  Each of these challenges reduces the likelihood that partners will use rare and tedious internet access to edit their maps at the office. Learning from Existing Systems Since arriving in June, our team assumed there was in fact a defined community of early warning and response (EW/R) actors that would find geolocated conflict data useful.  We have learned that there is a less cohesive EW/R community than we originally thought; rather, it seems like there are some loosely affiliated organizations and institutions that devote some attention to this field but very few that are entirely devoted to EW/R.  While there seem to be recognized responders in Liberia – the National Police, the Armed Forces, the UN – it is more difficult identifying early warning actors.  Even those organizations that have EW activities have divided their attention among many different projects and do not focus on EW per se.  This makes catering to these actors more difficult when they themselves are not easily identifiable. As a participant in Liberia’s Early Warning Working Group (EWWG), our team has been able to better understand who consider themselves members of the EW community.  The EWWG is evidence that there’s a growing interest among civil society organizations to share their knowledge with one another and to collectively contribute useful information to responders.  This group’s formation is an encouraging step towards a more defined EW community and their ability to contribute to conflict detection, mitigation, and prevention. It should however be mentioned that, while response actors are easier to identify, our original assumptions that response actors are capable of utilizing EW data and mobilizing quickly are not accurate.  It is quite clear that response actors in Liberia lack the coordination and the resources for effective response.  The UN remains the most effective responder, but rarely does the UN collaborate with national response actors and even less with early warning actors.  This lack of cooperation may make the formation of an early warning/response ecosystem more complex and challenging than we originally anticipated. In the course of attending EWWG meetings and workshops, our team has noticed that many members have predetermined metrics and indicators for early warning and conflict prevention.  This makes the Working Group - and more specifically our team’s efforts to compile data into a collective early warning mega-map – more challenging since participants have preexisting attachments to certain methodologies.  This project was partly reliant on the assumption that platform users have a shared idea of conflict and what unrest looks like; in reality, participants have different perspectives on these subjects, and these differences will influence our future efforts to compile data on shared platforms for early warning and for the election. In this current project’s proposal, some key assumptions were listed: actors will self-organize without external intervention, and engagement will be sustained.  Our team now realizes these assumptions are unrealistic.  Considering that most partners are currently reporting at a low frequency, we anticipate that partners’ use of the tool will need to be monitored and facilitated beyond December of this year.  Our timeframe was ambitious and the unexpected challenges have been numerous; however, our hope that this tool would be useful in Liberia is still alive and kicking.  It may be that our reach will be limited and our crowdsourcing bounded, but because of this platform there will still be more information about Liberia in the public domain.  As our team continues tweaking the tool, and our approaches to its application, we’ll no doubt have more lessons to share; in the meantime, we’d love to hear your thoughts, what you’ve learned, and any insights from which all of us out in the field could benefit.