GIS Analyst Interview Questions
& Model Answers, 2026

A datum defines the reference surface and origin used to express positions on the earth, while a projection is the mathematical transformation from that curved surface to a plane, each distorting area, shape, distance or direction differently. Mixing datasets on different datums shifts features by metres without any error message, and choosing a projection that distorts the property you are measuring produces wrong areas or distances. Say what you check first when two layers do not line up.

Start from the decision the analysis must support, then match the method: overlay for combining thematic conditions, proximity and buffering for distance relationships, network analysis where travel follows a route rather than a straight line, and surface or interpolation methods for continuous phenomena, being explicit about the assumptions each makes. Say why a straight-line buffer is wrong for a drive-time question and how often that mistake is made.

Check the lineage and source, the capture date and currency, the positional accuracy and scale it was captured at, the completeness and attribute reliability, and whether its resolution supports the decision being made. Say why scale matters — data digitised for regional mapping is not suitable for parcel-level decisions no matter how good it looks zoomed in. Document the limitations with the analysis rather than after someone acts on it.

Address matching against a reference dataset with a scoring threshold, producing rooftop, parcel, street interpolation or centroid results of very different accuracy. It fails on incomplete or non-standard addresses, new developments not yet in the reference data, rural addresses, and units within buildings, and interpolated results can be far from the actual location. Say why you keep the match score and type with the record — an analysis treating all geocodes as equal is misleading.

A clear question answered by the map, a classification method chosen for the data distribution rather than the default, colour that is perceptually ordered and accessible to colour-blind readers, normalised values rather than raw counts where area varies, and only the elements that support the message. Say why an unnormalised choropleth of counts is misleading — it is mostly a map of where people live.

Script the workflow so it is reproducible and version-controlled rather than clicking through the same steps, parameterise inputs, include data validation and logging so a silent input change is caught, and schedule where the output is needed regularly. Say why reproducibility matters beyond time saving — an analysis nobody can rerun cannot be audited or corrected when the underlying data changes.

Describe having the method, the data sources and the assumptions documented, reviewing the challenge honestly, correcting if warranted, and explaining the limitations plainly. Say how you present uncertainty. Spatial analysis carries an air of precision that the underlying data often does not justify, and analysts who present results without their limitations get caught out.

Talk about asking what decision the output will inform rather than what map they want, sketching an example early to test the direction, and iterating rather than delivering a finished product against a guess. Say what changed once they saw a draft. Most GIS rework comes from delivering exactly what was asked for rather than what was needed.

Good examples: a datum shift, duplicated or missing features, an attribute join on a non-unique key, a stale dataset presented as current, or geocodes at a default location. Describe how you found it and what you did about the analyses that had already used it. Say how you prevented recurrence. Silent data errors propagate widely because GIS outputs look authoritative.

Describe leading with the finding rather than the method, using a simple visual, being explicit about confidence, and answering the question actually asked. Say how you handled someone who wanted a stronger conclusion than the data supported. Maps are persuasive out of proportion to their evidence, and that carries a responsibility to state limits clearly.

Anchor to the occupation this role is reported under: the BLS OEWS May 2025 median for cartographers and photogrammetrists is $81,390 a year ($39.13/hr), with the top 10% above $126,950. Position by analysis depth, programming and database capability, domain knowledge in the employer's sector, and whether the role owns data management as well as analysis.

Programming and automation, spatial database and data engineering skills, remote sensing or imagery analysis where relevant, and domain expertise in a sector such as utilities, telecoms or environment. Ask whether the role is analysis and development or map production, since those are very different jobs with very different trajectories despite sharing a title.

Software and data access, training in programming and database skills, conference attendance, and hardware appropriate for large datasets. Ask how much of the role is producing routine map products, because a job that is largely repetitive cartography will not build the skills that move you up the range.