When visualizing national highway mileage across Chinese provinces in QGIS, I needed provincial capital markers without local data access. Using DeepSeek AI, I generated geographic layers directly in QGIS via Python scripting.


Visualization requiring provincial capitals

Methodology

1. Account Setup

Registered at DeepSeek and accessed the chat interface:

2. Prompt Engineering

Initial request:

"Provide coordinates for all Chinese provincial capitals and generate executable Python code for QGIS console display."

DeepSeek response included complete implementation:
python Capital city coordinates dictionary

capital_cities = { "北京": (116.4074, 39.9042), "天津": (117.1994, 39.0851), # ... (all 34 capitals) }

Layer creation and styling

layer = QgsVectorLayer("Point?crs=EPSG:4326", "Provincial Capitals", "memory") provider = layer.dataProvider() provider.addAttributes([ QgsField("city", QVariant.String), QgsField("longitude", QVariant.Double), QgsField("latitude", QVariant.Double) ]) ... (full implementation)

3. Execution

Copied code to QGIS Python Console:

Extended Functionality

Basemap Integration

Prompt:

"Generate code to load Tianditu satellite basemap in QGIS Python Console."

Solution (adjusted for working OSM alternative):
python OSM Satellite Basemap

from qgis.core import QgsRasterLayer, QgsProject osm_url = "https://tile.openstreetmap.org/{z}/{x}/{y}.png" osm_layer = QgsRasterLayer("type=xyz&url=" + osm_url, "OSM Satellite", "wms") QgsProject.instance().addMapLayer(osm_layer)

Voronoi Polygons

Prompt:

"Generate Voronoi polygons from point layer in QGIS Python Console."

Gemini-provided implementation:
python from qgis.core import * from qgis.processing import * point_layer = project.mapLayersByName("your_point_layer_name")[0] params = {'INPUT': point_layer, 'OUTPUT': 'memory:'} output_layer = processing.run('qgis:voronoipolygons', params)['OUTPUT'] project.addMapLayer(output_layer)

Performance Assessment

  1. DeepSeek Advantages:
    • Generated complete, executable QGIS solutions
    • Understood complex GIS requirements
    • Provided commented, structured code
  2. Limitations:
    • Server instability during peak times
    • Restricted advanced processing prompts
    • Tianditu implementation required debugging

This workflow demonstrates how AI assistants augment GIS workflows by rapidly generating executable geographic solutions without traditional data preparation.