Yesterday, someone in a group shared a map of China's 2024 birth rate, which looked quite interesting. However, this map was different from the usual ones—regions with lower birth rates were colored redder. Someone asked if it could be quickly recreated.

Two years ago, I might have been too lazy to bother because redrawing the map would require manually extracting and recording data from the map one by one, then importing it into QGIS or ArcGIS for data processing, and finally producing the map. But in 2025, the era of AI, this task is a breeze. Today, I'll share a set of AI-based rapid map production methods. Next time you encounter such a demand, just go for it!

Above: Image sourced from the official account @EPC1314

AI Data Extraction

In the past, how would we extract this data? Would we open an Excel spreadsheet and enter it one by one? Now, we can directly use AI with the following prompt:

Read the current image, organize the names of Chinese provinces and their corresponding numbers, and compile the results into a CSV file.

The output is as follows:

It is important to note that the downloaded data must be verified. I compared the output results from Gemini and ChatGPT, and without exception, both had errors. ChatGPT's output had slightly fewer errors, but it's still advisable to compare results from multiple AI tools.

Importing into QGIS

Once the data is downloaded, the next step is map production using GIS software. This part is relatively straightforward. Open QGIS and load the Chinese map shared in our previous article, "「GIS Data」2024 National Standard Vector Map (Precise to County Level) Approval Number: GS(2024) 0650", then project it using Albers.

Note: QGIS does not natively include the Albers projection for China, so you need to define it yourself. For the method, refer to the article: "QGIS Add Custom Projection".

Import the CSV file from the first step into QGIS.

This CSV file contains geometric information, so be sure to check the corresponding option.

Next, right-click on the map layer, select Properties, and find the Joins section.

Set the fields from the CSV file and the target fields accordingly. Finally, set the symbology by right-clicking on the layer, selecting Properties, finding the Symbology section, and configuring the symbols.

Find a similar color ramp in QGIS's default symbol system. The "turbo" ramp I found also uses red for smaller values and green for larger ones. You can manually reverse it if needed. The final result is as follows:

Map Production in ArcGIS

The operation in ArcGIS is similar. Unfortunately, my Windows computer is broken and not yet repaired, so this section will be left blank for now and updated later.

Data Download

If you want to try this out and don't want to verify the data one by one, I'll share the data I generated using AI. Feel free to download it if needed.

china_province_birth_rate_2024.7z

Link: https://pan.baidu.com/s/1EroeRN7DeJbQOo9EzRNvvw Extraction code: mpum

Unzip password: malagis.com

Final Thoughts

From seeing the map to producing the final result, it took me less than 20 minutes. AI has, in some ways, truly changed the way we create maps in GIS. In the past, I wouldn't have bothered with such tasks (especially since there's no pay), but with AI, I now have time to tackle these interesting little projects.

Finally, always remember:

Do not upload classified data to foreign AI tools, nor to domestic ones. Classified data must not go online; online data must not be classified!!!!