This is the second of a two-part post exploring questions on why and how to map in support of peacebuilding and protection outcomes. A few weeks ago, I blogged about the Iraq Dispute Mapping and Early Warning system that Mercy Corps has recently put in place. Before deciding on a mapping platform, the design process for the system sought to understand its primary users, the programmatic function of data collection and the users’ analysis needs. In designing the Mercy Corps Libya Protection Mapping System, we followed a similar process.
Understand the user: the Mercy Corps Libya Protection Team
The Mercy Corps Libya Protection Program gathers detailed data on demographics, socio-economic conditions and conflict indicators from the IDP sites where Mercy Corps works (currently about 75, although that number fluctuates). Data is gathered during regular (monthly or bi-monthly) protection monitoring site visits by Mercy Corps staff, who interview key stakeholders at the site. Although information is gathered only from these key stakeholders (not through individual or family level randomized surveys), it constitutes the richest, most accurate and most up to date source of data on IDP populations in Libya.
Identify the main function: internal reporting and program design
The Mercy Corps team already reports regularly on the rich data it collects. Internal reports are mostly aimed at donors and other actors working with IDPs, including the Government’s aid agency (LibAid). One function of the system is to systematize the reporting process, making some tasks automatic. The other critical function is to allow for easy comparisons over time and across geographic areas, so that conclusions can be drawn about emerging patterns and needs. This will turn the reports from narratives of the situation into tools for designing future programming.
Displaced populations are a sensitive issue in Libya, so balancing ethical considerations with functionality is important. Certain populations are associated with military forces that defended the Ghadafi regime and victimized (at times even attacked) as a result. This is a particular problem with the Tawerghan identity group. It is therefore imperative to take into account do no harm principles with respect to any data collection or processing about IDPs. It seems unlikely that the data in the system can be directly used for harm, that is to target IDPs. The location of IDP camps is not hidden from the public. If anything, it may raise public awareness of IDP issues in Libya. However, sharing data can have a secondary harmful effect: it could change the perception IDPs have of the Mercy Corps teams. Some team members believe that if data is shared, then IDPs will feel their trust has been violated.
Decide on data collection functions: web based long form
Prior to introduction of this system, the Mercy Corps Protection Team already collected large amounts of information. The data collection process was not fully standardized, making comparisons at times difficult. Some data was uploaded to summary Google spreadsheets, but much remained in paper forms. The team agreed that a standard, online form would be a helpful tool to consolidating data inputs.
In designing this form, two methodological issues came up. First, the initial set of questions proposed exceeded 256 (which for all you non-geeks is the maximum number of columns a standard spreadsheet can hold). The team also agreed that trying to go for so much quantity of information might dilute the quality of what was collected due to time constraints. Second, the unit of analysis for the questionnaires is the site, not individuals. This required some rewording of questions to avoid misleading answers. For example, it doesn’t make sense to ask what the preferred solution to displacement is for IDPs, since there is likely to be a preference for all solutions by at least one IDP at the site. The question was rephrased to indicate the top preferred solution at the site, so that the data can provide a sense of priorities across sites and across time.
Understand what analysis is needed: descriptive summaries
The Protection Team requires a flexible tool that will allow them to translate collected data into inputs to program design, implementation and monitoring. The team considers geographic visualization important to highlight differences across regions. Equally important is thematic analysis of indicators, using charts and numerical summaries to highlight issues or groups that require special attention. Finally, visualizing change over time (on a map or in graphs) is important to tracking progress and identifying emerging trends.
Choose your mapping technology: Google Crisis Map
With these requirements in mind, two possible technology solutions were explored: Ushahidi and a combination of Google products. The Google solution was deemed preferable for two main reasons. First, Ushahidi works well with few indicators over many locations, but is not well suited to the opposite situation (few locations and many indicators). Second, the graphic and summary analysis capacities of Ushahidi are very limited, and manipulation of downloaded data is not straightforward. Google offers more complete, inbuilt graphical and summary tools.
The proposed system design is outlined in the diagram below. The system utilizes four Google components to deliver all the functionality needed. Users enter data to the system through a Google Form. The form is designed to make data entry fast and systematic, by providing multiple choice answers for the majority of questions. It is password protected and available only to users with a Mercy Corps Libya email address. Once a user finishes filling out a report form, s/he receives a confirmation email that also contains a copy of the filled out form. The email provides a link where the user can correct / add data.
Each submitted report form is automatically added to a Google Fusion Table. The table is available only to the system administrators, who can open each record to review it for quality and then approve it. This master table also gives the system administrator tools to filter and aggregate by any indicator, viewing them in tables or graphs. The system administrator can also create thematic fusion table views. These “views” are selected subsets of indicators for a particular theme. The views are merged with a geolocation reference sheet of IDP sites, automatically producing a map layer. All the thematic fusion table views are aggregated on the platform that powers Google Crisis Maps to create an interactive, online map. Each table view constitutes a “layer” in this map, which can be checked on and off for viewing in the mapping interface.
Finally, the online map is housed on a Google Site. This site is password protected and made available to a target audience, including Mercy Corps staff, donors and partner staff. It contains thematic pages for topics of interest, showing a selection of graphs made with the master fusion table. It also contains summary ‘cards’ for each IDP site by month, showing a selection of the questions in the data collection form.
Mapping and beyond: Mercy Corps and tech
As with the Iraq system, the Libya Protection Mapping System is the result of a design process that tailors available, user-friendly, free mapping technologies to programmatic needs. It is a system with a concrete purpose and delivers much more than red dots. The main take-away from these two projects goes beyond mapping, and points to what should be our best practices in engaging tech for programming in conflict settings. Too many systems are designed out of a desire to try a particular tech tool. The principle of designing from the user need, simple as it may seem, is not always followed. It also takes longer, a bit more trial and error, which is why it’s exciting to work with teams that are willing to invest time in thinking through these questions. Mercy Corps recently approached me to run a workshop on peacebuilding and technology – so watch this space for more tech work they may start exploring soon.