When Maps Become Evidence: Using Spatial Analysis to Strengthen Evaluation

By Frank Salet

Most evaluations include maps, but they rarely play a meaningful role in analysis. They are used to show where activities took place, helping the reader orient themselves before moving on to findings that are treated as separate.

In stable settings, separation is often acceptable. In fragile or hard-to-access environments, as we have argued, it becomes a limitation—because when direct observation is uneven or impossible, location is not just context. It is part of how we assess whether findings are credible.

The Limits of Mapping as Illustration

Maps are often produced at the end of an evaluation process, reflecting conclusions that have already been formed. As a result, they do not test those conclusions; they simply illustrate them.

This approach struggles in contexts where access varies significantly, where some areas cannot be visited, and where data is collected through a mix of fieldwork, remote methods, and secondary sources. In these settings, the issue is not a lack of data, but how that data is interpreted across space.

Without integrating geographic context into the analysis, it becomes difficult to distinguish between what has been verified and what has simply not been examined.

Starting with the Map—But Not Stopping There

In Apricity’s recent Yemen evaluation, maps were not treated as outputs, but as part of how the evaluation was designed.

The first step was to map where projects were implemented, and then immediately layer that with areas of state and non-state control, population distribution, and access constraints. This combination did not just provide context—it shaped the evaluation strategy itself.

This step clarified where fieldwork would be feasible, where it would be partial, and where it would not be possible. Those distinctions directly informed how data was collected: where third-party teams could visit, where mobile contributors could operate, and where no direct verification would be attempted due to security and duty-of-care considerations.

Used this way, maps move from simply describing decisions to shaping them—functioning as part of the evaluation evidence base from the outset. 

Layering Data to Test What You See

As an evaluation progresses, spatial analysis can become a way to bring the different evidence streams we described in our last piece together and assess how they relate.

In Yemen, project locations were layered with geo-referenced surveys, site visits, community-sourced observations, and satellite imagery showing physical changes over time. This made it possible to examine each location through multiple lenses and compare how different data sources aligned.

In some cases, the evidence converged clearly. Infrastructure reported in project documents could be observed in satellite imagery, located by field teams, and confirmed through community reporting, which strengthened confidence in those findings.

In other cases, the picture was more complex. Some sites appeared structurally intact in imagery but showed limited functionality or maintenance in field observations. In a few instances, reported outputs could not be fully confirmed or appeared to have deteriorated over time when viewed across sources.

Rather than treating these differences as inconsistencies to resolve, they became part of the analysis, with spatial context helping to assess where findings were strong, where they were uncertain, and why.

Illustrative: Annotating infrastructure, access constraints, and service gaps directly onto satellite and field maps as part of evaluation planning and evidence review. 

Context Shapes What Results Mean

Spatial context also played a central role in interpreting findings.

Mapping project locations alongside areas of government and armed group control in Yemen made it clear that some regions could not be evaluated directly. Activities in those areas were not excluded, but they were interpreted differently, with the absence of verification understood as a result of access constraints rather than a simple data gap.

This distinction matters because it allows evaluators to be transparent about where evidence is stronger or weaker without penalizing programs for operating in difficult environments.

Spatial patterns also helped explain differences in outcomes. For example, in Yemen, infrastructure performance varied across locations, and this variation often aligned with system type and local conditions. Piped systems with fee collection tended to remain functional and well-maintained, while systems without revenue mechanisms were more likely to deteriorate over time.

Seen spatially, these are not isolated findings but part of a broader pattern that links context, design, and sustainability.

Seeing Gaps More Clearly

A spatial lens also makes it easier to identify what is not happening.

When project footprints are viewed alongside population distribution and access constraints, gaps in coverage become visible. These may include areas with high need but limited intervention, or regions that are consistently excluded due to security or logistical barriers.

These patterns are rarely random. They reflect the operational realities of working in fragile contexts, and making them visible allows evaluators to assess how those constraints shape both implementation and outcomes.

From Maps to Exploration

Maps also played a role in how findings were communicated.

Static maps provided an overview of project locations and reported results, while interactive tools — such as the StoryMap developed alongside the evaluation — allowed users to explore individual sites, view evidence, and understand how different data sources related to each other.

This approach does not replace interpretation, but it makes the underlying evidence more transparent and allows others to engage with it more directly.

What this Changes in Practice

Using spatial analysis in this way shifts the role of maps from presentation to analysis. They become part of how evaluations are designed, how data is interpreted, and how conclusions are tested.

It also brings context into the center of the evaluation, making access constraints, geographic distribution, and visibility part of how evidence is assessed rather than treated as secondary considerations.

At the same time, spatial analysis has limits. It relies on data quality, requires careful interpretation, and is most effective when used alongside other forms of evidence rather than on its own.

For practitioners, this means treating location as part of the data from the outset. For donors, it means looking beyond outputs to understand how geography shapes both implementation and evidence.

Maps do more than show what happened; they help determine whether it makes sense.

In our next piece, we discuss one such complementary tool—satellite imagery—and what it actually requires to use well.