This page needs a newer browser to run its tools. You can still read everything here.

GIS & Geospatial Data (QGIS, GeoPandas, PostGIS) Module 1

Foundations: spatial questions and data models interview questions

Interview questions on “Foundations: spatial questions and data models”: warm-ups, then what each harder question tests and how a good answer starts.

9 questions: 3 warm-ups and 6 with model answers · Levels: Fresher to senior

Warm-up questions

Short questions with short answers, each answered in full.

Fresher

What is the difference between a map and a GIS?

Answered in the lesson What GIS is and the questions it answers
Fresher

When would you store data as a raster rather than as vectors?

Answered in the lesson Vector and raster data models
Fresher

What is the difference between the precision and the accuracy of a coordinate?

Answered in the lesson Latitude, longitude and coordinate precision

Questions with model answers

What each question tests, the first step of a good answer and the follow-ups to expect.

Fresher to mid

A health department asks whether its clinics are "in the right places". How do you turn that into spatial analysis, which data and methods do you use, and how do you check the result before anyone acts on it?
What it tests

Framing a vague request as spatial questions, choosing data and units that fit the operation, and checking a result in two independent ways.

First step

Ask what "right" means and for whom: a distance or travel-time limit, the people who count (every household, or families with young children), and the decision the answer feeds. Then name the shapes: proximity for coverage, a network if people travel by road, containment to report by ward.
Follow-ups
- The department says "within 15 minutes of travel". What changes in your analysis?
- Two tools give different coverage counts for the same data. How do you find the cause?
- The department can open one more clinic. How would you choose where?
From the lesson What GIS is and the questions it answers
Model answer, Pro. Model answers open with any pass; prices are in the notice on this page.
Opening the model answers with your pass…
Model answers open with any MySmartCoPilot pass.
Pro pass · 1 week₹79$2.99
30 days₹179$5.99
1 year₹1,699$45
One payment; a pass never renews. The price shown is the total you pay: no fees or charges are added at checkout.
Passes can’t be bought right now, so the model answers can’t be opened at the moment. If you have a pass, enter your licence key.
Choose a pass I have a licence key
It also opens every other in-depth part in Learn and every Pro tool. Sales are final, except where the law gives you a right to withdraw, for example in the EU and the UK (refund policy).
In the answer:
- Model answer
- Follow-up answers
- Common mistakes
- Rubric
- Variants
- about 4 minutes
Mid to senior

A model predicts noise levels from 400 sensor readings across a town. Tested on a random fifth of the sensors its error is small, but in a new district it does badly. Explain what happened, and how you would evaluate the model properly.
What it tests

Recognising spatial autocorrelation behind an optimistic score, explaining why random splits leak information, and designing a validation that matches how the model will be used.

First step

Measure how far each test sensor is from its nearest training sensor. If the median is a few hundred metres while the new district is kilometres from any sensor, the random test answered an easier question than the real one, because nearby readings are alike.
Follow-ups
- How do you choose the size and shape of the blocks you hold out?
- Can a spatially blocked test be too pessimistic? When?
- Which checks show that spatial dependence is left in the model's errors?
From the lesson What GIS is and the questions it answers
Model answer, Pro.
In the answer:
- Model answer
- Follow-up answers
- Common mistakes
- Rubric
- Variants
- about 4 minutes
Fresher to mid

From the same survey, one analyst reports a lake's area as 69.4 ha from its shoreline polygon, and another as 68.0 ha from a 100 m land-and-water raster. Which do you trust, why do the numbers differ, and what do you tell the team?
What it tests

Understanding rasterisation rules and cell size, knowing which measurements converge as cells shrink, and choosing the model that fits the question.

First step

Ask how the raster was made: the cell size and the rule for cells on the shore. With 100 m cells and a cell-centre rule, every shore cell is all water or all land, so a difference of about 2 % is expected, and any feature narrower than a cell can vanish.
Follow-ups
- The team also needs the length of the shoreline. Which number do you give?
- Would counting every cell the lake touches give a better area?
- What cell size would you choose to keep a 40 m inlet?
From the lesson Vector and raster data models
Model answer, Pro.
In the answer:
- Model answer
- Follow-up answers
- Common mistakes
- Rubric
- Variants
- about 3 minutes
Mid to senior

A council wants to charge a fee to every building inside a town boundary that was digitised from 1:50,000 maps and states a horizontal error of 12.5 m RMSE. What risks do you raise, and how would you decide the buildings near the line?
What it tests

Turning map scale and a stated RMSE into distances on the ground, quantifying the chance of a wrong side, and designing a fair decision process.

First step

Turn the numbers into metres: at 1:50,000 a 0.25 mm line is 12.5 m wide, and with 12.5 m RMSE about 95 % of errors fall within 21.6 m. Buildings well inside or outside are safe; those within a few tens of metres of the line need better evidence.
Follow-ups
- How many buildings would you expect in the uncertain band, and how do you find them?
- Which better data would settle the cases near the line?
- The council insists on using the layer as it is. What do you put in writing?
From the lesson Vector and raster data models
Model answer, Pro.
In the answer:
- Model answer
- Follow-up answers
- Common mistakes
- Rubric
- Variants
- about 3 minutes
Fresher to mid

About 3 % of the shops on your company's map appear in the sea north of Norway, though every shop is in India. The data comes from several partner feeds. How do you find the cause, fix the data and stop it from happening again?
What it tests

Recognising swapped latitude and longitude from where the points land, tracing it to a source, repairing it safely and adding checks that catch it.

First step

Look at where the wrong points are: if a shop at 18.52° N, 73.86° E shows at 73.86° N, 18.52° E, the numbers are swapped. Group the bad rows by partner feed and by the code path that read them, because one feed or one conversion usually explains them all.
Follow-ups
- How do you repair the 3 % without breaking the 97 % that are right?
- Which automatic check would you add, and where in the pipeline?
- The company expands to a region from 0 to 10° N and 5° W to 10° E. Does the same check still work?
From the lesson Latitude, longitude and coordinate precision
Model answer, Pro.
In the answer:
- Model answer
- Follow-up answers
- Common mistakes
- Rubric
- Variants
- about 4 minutes
Mid to senior

A delivery app stores customer locations with 15 decimals and sends them to phones as GeoJSON. A colleague proposes rounding to 5 decimals everywhere. How do you decide how many decimals to keep for storage, for the API and for display?
What it tests

Turning decimal places into ground distance, separating precision from accuracy, and weighing payload size against what each consumer of the data needs.

First step

Work out what the decimals mean on the ground: the fifth is about 1.1 m and the sixth about 11 cm of latitude. Then ask how accurate the positions are, typically a few metres for phone fixes, and what each consumer does with them: routing a driver, drawing a pin, or settling a dispute.
Follow-ups
- Is there any reason to keep more decimals in the database than in the API?
- How much does the API payload shrink, and does it matter?
- A few customers are surveyed to the centimetre. What changes?
From the lesson Latitude, longitude and coordinate precision
Model answer, Pro.
In the answer:
- Model answer
- Follow-up answers
- Common mistakes
- Rubric
- Variants
- about 3 minutes

All GIS & Geospatial Data (QGIS, GeoPandas, PostGIS) interview questions · The GIS & Geospatial Data (QGIS, GeoPandas, PostGIS) track

Warm-up questions

What is the difference between a map and a GIS?

When would you store data as a raster rather than as vectors?

What is the difference between the precision and the accuracy of a coordinate?

Questions with model answers

A health department asks whether its clinics are "in the right places". How do you turn that into spatial analysis, which data and methods do you use, and how do you check the result before anyone acts on it?

A model predicts noise levels from 400 sensor readings across a town. Tested on a random fifth of the sensors its error is small, but in a new district it does badly. Explain what happened, and how you would evaluate the model properly.

From the same survey, one analyst reports a lake's area as 69.4 ha from its shoreline polygon, and another as 68.0 ha from a 100 m land-and-water raster. Which do you trust, why do the numbers differ, and what do you tell the team?

A council wants to charge a fee to every building inside a town boundary that was digitised from 1:50,000 maps and states a horizontal error of 12.5 m RMSE. What risks do you raise, and how would you decide the buildings near the line?

About 3 % of the shops on your company's map appear in the sea north of Norway, though every shop is in India. The data comes from several partner feeds. How do you find the cause, fix the data and stop it from happening again?

A delivery app stores customer locations with 15 decimals and sends them to phones as GeoJSON. A colleague proposes rounding to 5 decimals everywhere. How do you decide how many decimals to keep for storage, for the API and for display?