GIS Interview Questions¶

The questions below are drawn from real entry-level and analyst-level GIS interviews. They fall into 5 categories — technical, conceptual, scenario, portfolio walk-through, and behavioral.

For each, we list what they're really asking and a sample answer.

Conceptual¶

1. What is GIS?¶

What they're really asking: Can you describe GIS to a non-GIS person?

Sample answer:

"A GIS is software for working with location-based data. We capture features as points, lines, or polygons, attach attributes to them, and analyze the spatial relationships — like which neighborhoods are near a hospital, or which parcels are in a flood zone. The output is usually a map, a dashboard, or a tabular result that informs a decision."

2. What's the difference between GCS and PCS?¶

Sample answer:

"A Geographic Coordinate System uses latitude and longitude on a 3D model of the earth. A Projected Coordinate System flattens the earth onto a 2D plane and uses x, y in linear units like feet or meters — better for measuring area and distance. The right projection depends on the area and the analysis."

→ See Coordinate Systems.

3. What's the difference between vector and raster?¶

Sample answer:

"Vectors model discrete features — roads, parcels, points of interest — as points, lines, and polygons with attribute tables. Rasters model continuous phenomena — elevation, temperature, satellite imagery — as a grid of cells with values. Vectors are precise but heavy for continuous data. Rasters are great for surfaces but blocky for crisp boundaries."

4. What's the difference between Buffer, Clip, and Intersect?¶

Sample answer:

"Buffer creates a zone of a specified distance around a feature. Clip uses one layer's polygons as a cookie cutter on another. Intersect returns only the overlapping geometry between two layers and combines their attributes. I use Buffer for proximity, Clip for restricting analyses to an area, and Intersect for spatial overlay."

→ See Buffer vs Clip vs Intersect.

5. Spatial join vs attribute join?¶

Sample answer:

"An attribute join uses a common field — like a county FIPS — to link tables. A spatial join uses location: 'which polygon is each point inside?' or 'how many points fall in each polygon?'. Spatial joins are essential when you don't have a shared key."

6. What makes a good map?¶

Sample answer:

"Clear hierarchy — the eye knows what's most important. A focused message — answers one question. The right projection. Clean typography. A title, legend, scale, north arrow, and attribution. Color choices that match the data type — sequential for ordered, divergent for around-zero, qualitative for categories."

Technical¶

7. How would you find all schools within 1 mile of a park?¶

Sample answer:

"I'd buffer parks by 1 mile (in a projected CRS), then run Select by Location on the schools layer using 'intersect' with the buffer. Or use a spatial join with a 1-mile distance match. Quick QA check that I get the expected order of magnitude before publishing."

8. How would you compute the population in a flood zone when populations are tied to census tracts?¶

Sample answer:

"Two approaches: a quick areal-weighted method, where I compute the percent of each tract's area inside the flood zone using Tabulate Intersection, then multiply by the tract population. A higher-fidelity method uses dasymetric mapping, where I weight by housing units or impervious surface within each tract. The first is good for a starting estimate; the second is what you'd defend to a stakeholder."

9. You join two tables and most rows are null. What's wrong?¶

Sample answer:

"Almost always a key mismatch — either type (text vs number), formatting ('06037' vs '6037'), trailing whitespace, or case. I check field types, sample 10 keys side-by-side, look for leading zeros and casing, and use a left outer join to see the unmatched records."

→ See How to Clean Data Before Joining.

10. What's an example SQL query you've written in ArcGIS Pro?¶

Sample answer:

"To find populated census tracts above 65: POP_TOTAL > 0 AND PCT_OVER_65 > 25. To handle multi-FIPS: STATE_FIPS IN ('06','12','48'). To exclude null incomes I always combine with AND MED_INC IS NOT NULL."

→ See SQL for GIS.

11. What is a spatial index, and when do you need one?¶

Sample answer:

"A spatial index speeds up queries that filter by geometry — overlap, intersect, distance — by pre-organizing features into spatial bins. Most GIS layers have one by default; if a query is slow, you may need to rebuild or add one. In databases like PostGIS, I'd create a GiST index on the geometry column."

12. What is topology and when does it matter?¶

Sample answer:

"Topology is a set of rules describing how features relate: parcels can't overlap; roads must be connected at intersections; counties tile a state without gaps. It matters when the geometric correctness of features matters — utilities, parcels, transportation. ArcGIS Pro has built-in topology rules with validation tools."

Scenario¶

13. A planner asks you for "a map of where new affordable housing should go." How do you respond?¶

What they're really asking: Do you ask scoping questions before you start clicking?

Sample answer:

"I'd ask: what counts as 'affordable housing' here — definition and density? What's the study area? What constraints — zoning, environmental, parcel ownership? What positive criteria — near transit, near jobs, near schools? Then I'd propose a multi-criteria suitability analysis with explicit weights, agree on the weights with the planner, and produce a draft. Showing iteration to the planner early avoids re-doing analysis."

→ See Site Suitability project.

14. Your analysis returns 0 rows. What do you check?¶

Sample answer:

"I sanity-check the inputs first: did the layers load, are they in the right CRS, are there records at all? Then I look at my SQL — did I add a typo, an unmet AND, a wrong field name? I might temporarily relax the WHERE clause to confirm rows exist, then add conditions back one by one to find which one zeroed me out."

15. The dataset has 1.2M rows and the geoprocessing tool runs forever. What do you do?¶

Sample answer:

"Run on a small subset first — 1k rows — to validate the workflow. Add a spatial index. Use environment settings (extent, mask) to limit work. Switch to in-memory workspaces for speed. If it's still slow, consider a database (PostGIS) or chunking with arcpy/geopandas."

Portfolio walk-through¶

16. "Walk me through one of your projects."¶

What they're really asking: Can you tell a story? Did you actually understand it?

Use a STAR-like structure:

• Situation — the question / problem
• Task — what you decided to deliver
• Action — your steps and decisions
• Result — outputs, numbers, who it served

"I built a transit desert analysis for Austin. The question was: what neighborhoods have the worst transit access compared to where people live? I pulled GTFS data and ACS at the tract level, generated 0.5-mile service area buffers around stops, computed percent of tract area covered, joined it with population, and built a Story Map and dashboard. I found 12% of the metro population live in tracts with poor service. The biggest decision was using actual transit network walks instead of straight-line distance, because of barriers like highways."

17. "What was the hardest part of that project?"¶

Sample answer:

"Cleaning GTFS — feed quality varied by agency. I built a small Python script to QA the stop count and route count per provider, and ended up dropping one provider's service because of broken stop sequences."

18. "If you could redo it, what would you change?"¶

Show humility, growth, awareness.

Sample answer:

"I'd build it on a network analysis instead of straight-line buffers. I'd also add a temporal dimension — service availability at different times of day. The current version is essentially a daytime snapshot."

Behavioral¶

19. "Tell me about a time you handled feedback or criticism."¶

Sample answer (tell a story):

"On my food desert project, I showed a draft to a friend in public health. She pointed out I was using straight-line distance, but the realistic walk distance is much longer because of highways. I rebuilt the analysis using network distance and the desert classification changed for ~15% of tracts. I document my methodology more carefully now."

20. "Why GIS?"¶

Sample answer:

"I love that GIS sits at the intersection of data and place. Real questions — where do people live, where can transit reach them, where is risk concentrated — have spatial answers. I've found I can build evidence-based maps that change a planner's or a stakeholder's mind in a way a spreadsheet never could."

Questions YOU should ask them¶

End every interview with thoughtful questions:

• "What does success look like in this role at 6 months?"
• "What's the typical project lifecycle here — solo, paired, or team-based?"
• "What's the data infrastructure — file geodatabases, enterprise, or both?"
• "What does the team value most about its analysts?"
• "Where do most analysts go from this role — promotion paths?"
• "What's a project you wish was prioritized that you haven't been able to do?"

Asking questions signals seniority and curiosity.

→ Next: Common GIS Job Tools.