Making Maps using AI Thoughtfully

Introduction

Making maps is easier than ever

Let’s see how easy it is

This month on the internet (feat. Claude)

Is clustering the right choice?

Should the creator review their creation?

Can a map hallucinate the existence of places?

AI is great at creating convincing outputs that look technically accurate on the surface

Under the surface, there could be 0 substance

The data

Knowing your data sources

Which of these areas contains 15 million people?

GBA (Municipal Corporation)

Bengaluru Urban (District)

Bengaluru Metropolitan Region Development Authority

Knowing the number is not enough, the full context of the source data is essential to use it correctly

Understanding why/how the data was collected

Is the sensor intentionally calibrated incorrectly?

Knowing how and why the data was collected is essential to understand how to use the data

Awareness of regional quirks

How many cities are there in China?

The basemap

Is Cubbon Park a:

Park
Garden
Urban Forest

What does your map show it as?

What is the name of this road?

The basemap does not appear out of thin air.

The basemap is a combination of many decisions by many people

Sometimes you need to understand why those decisions were made

Let’s check back on Claude

Closing Thoughts

Using AI is fine, but first think for yourself about how you would approach the problem yourself.

Put in the effort to validate and review the AI output, like how you would review a colleague.

It’s easier than ever to make maps. It’s also easier than ever to validate maps. Time saved on making, can be spent on validating.