Contextual Q&A with GeoAgent¶

GeoAgent can answer general questions using its built-in LLM — from earth science topics and natural disaster impacts to programming questions and everyday queries.

When a query asks for information or explanation rather than data retrieval or visualization, it is automatically routed to the ContextAgent for a direct LLM answer.

In [ ]:

# %pip install geoagent

# %pip install geoagent

In [ ]:

from geoagent import GeoAgent

agent = GeoAgent()

from geoagent import GeoAgent agent = GeoAgent()

1. Earth Science Questions¶

Ask about natural phenomena, environmental topics, and earth observation science.

In [ ]:

result = agent.chat("What is NDVI and how is it calculated?")
print(f"Intent: {result.plan.intent}")
print(f"Success: {result.success}")
print()
if result.answer_text:
    print(result.answer_text)

result = agent.chat("What is NDVI and how is it calculated?") print(f"Intent: {result.plan.intent}") print(f"Success: {result.success}") print() if result.answer_text: print(result.answer_text)

In [ ]:

result = agent.chat("How do synthetic aperture radar (SAR) satellites work?")
print(f"Intent: {result.plan.intent}")
print()
if result.answer_text:
    print(result.answer_text)

result = agent.chat("How do synthetic aperture radar (SAR) satellites work?") print(f"Intent: {result.plan.intent}") print() if result.answer_text: print(result.answer_text)

In [ ]:

result = agent.chat("Explain the difference between Landsat and Sentinel-2 satellites")
print(f"Intent: {result.plan.intent}")
print()
if result.answer_text:
    print(result.answer_text)

result = agent.chat("Explain the difference between Landsat and Sentinel-2 satellites") print(f"Intent: {result.plan.intent}") print() if result.answer_text: print(result.answer_text)

2. Natural Disaster Impact Questions¶

Ask about the impact of specific natural disasters — hurricanes, wildfires, floods, and more.

In [ ]:

result = agent.chat("How was NYC impacted by Hurricane Sandy?")
print(f"Intent: {result.plan.intent}")
print()
if result.answer_text:
    print(result.answer_text)

result = agent.chat("How was NYC impacted by Hurricane Sandy?") print(f"Intent: {result.plan.intent}") print() if result.answer_text: print(result.answer_text)

In [ ]:

result = agent.chat("What was the impact of Hurricane Florence 2018 in North Carolina?")
print(f"Intent: {result.plan.intent}")
print()
if result.answer_text:
    print(result.answer_text)

result = agent.chat("What was the impact of Hurricane Florence 2018 in North Carolina?") print(f"Intent: {result.plan.intent}") print() if result.answer_text: print(result.answer_text)

In [ ]:

result = agent.chat(
    "How did vegetation recover after flooding in Missouri River valley 2023?"
)
print(f"Intent: {result.plan.intent}")
print()
if result.answer_text:
    print(result.answer_text)

result = agent.chat( "How did vegetation recover after flooding in Missouri River valley 2023?" ) print(f"Intent: {result.plan.intent}") print() if result.answer_text: print(result.answer_text)

3. Climate and Environmental Trends¶

In [ ]:

result = agent.chat(
    "What are the long-term climate trends affecting Pacific Northwest forests?"
)
print(f"Intent: {result.plan.intent}")
print()
if result.answer_text:
    print(result.answer_text)

result = agent.chat( "What are the long-term climate trends affecting Pacific Northwest forests?" ) print(f"Intent: {result.plan.intent}") print() if result.answer_text: print(result.answer_text)

In [ ]:

result = agent.chat(
    "Explain the correlation between El Niño events and wildfire patterns"
)
print(f"Intent: {result.plan.intent}")
print()
if result.answer_text:
    print(result.answer_text)

result = agent.chat( "Explain the correlation between El Niño events and wildfire patterns" ) print(f"Intent: {result.plan.intent}") print() if result.answer_text: print(result.answer_text)

In [ ]:

result = agent.chat(
    "What is urban heat island effect and which cities are most affected?"
)
print(f"Intent: {result.plan.intent}")
print()
if result.answer_text:
    print(result.answer_text)

result = agent.chat( "What is urban heat island effect and which cities are most affected?" ) print(f"Intent: {result.plan.intent}") print() if result.answer_text: print(result.answer_text)

4. General Questions¶

GeoAgent can also handle general-purpose questions — it's not limited to earth science.

In [ ]:

result = agent.chat("What is the STAC specification and why is it important?")
print(f"Intent: {result.plan.intent}")
print()
if result.answer_text:
    print(result.answer_text)

result = agent.chat("What is the STAC specification and why is it important?") print(f"Intent: {result.plan.intent}") print() if result.answer_text: print(result.answer_text)

In [ ]:

result = agent.chat("What are Cloud Optimized GeoTIFFs and why are they useful?")
print(f"Intent: {result.plan.intent}")
print()
if result.answer_text:
    print(result.answer_text)

result = agent.chat("What are Cloud Optimized GeoTIFFs and why are they useful?") print(f"Intent: {result.plan.intent}") print() if result.answer_text: print(result.answer_text)

In [ ]:

result = agent.chat("What is the difference between raster and vector data in GIS?")
print(f"Intent: {result.plan.intent}")
print()
if result.answer_text:
    print(result.answer_text)

result = agent.chat("What is the difference between raster and vector data in GIS?") print(f"Intent: {result.plan.intent}") print() if result.answer_text: print(result.answer_text)

In [ ]:

result = agent.chat("What Python libraries are commonly used for geospatial analysis?")
print(f"Intent: {result.plan.intent}")
print()
if result.answer_text:
    print(result.answer_text)

result = agent.chat("What Python libraries are commonly used for geospatial analysis?") print(f"Intent: {result.plan.intent}") print() if result.answer_text: print(result.answer_text)

5. Mixed Queries — Data + Explanation¶

Some queries combine data visualization with contextual information. GeoAgent handles both in one response.

In [ ]:

# This should trigger data visualization
result = agent.chat(
    "Show me Sentinel-2 imagery of the Amazon rainforest in August 2024"
)
print(f"Intent: {result.plan.intent}")
print(f"Items: {result.data.total_items if result.data else 0}")
print(f"Has answer_text: {result.answer_text is not None}")
result.map

# This should trigger data visualization result = agent.chat( "Show me Sentinel-2 imagery of the Amazon rainforest in August 2024" ) print(f"Intent: {result.plan.intent}") print(f"Items: {result.data.total_items if result.data else 0}") print(f"Has answer_text: {result.answer_text is not None}") result.map

In [ ]:

# This should trigger explanation
result = agent.chat("What causes deforestation in the Amazon rainforest?")
print(f"Intent: {result.plan.intent}")
print(f"Has answer_text: {result.answer_text is not None}")
print()
if result.answer_text:
    print(result.answer_text)

# This should trigger explanation result = agent.chat("What causes deforestation in the Amazon rainforest?") print(f"Intent: {result.plan.intent}") print(f"Has answer_text: {result.answer_text is not None}") print() if result.answer_text: print(result.answer_text)

---

How It Works¶

GeoAgent's Planner classifies queries into intents:

Intent	Routing	Example
search	Data pipeline → STAC search	"Find Sentinel-2 images of Tokyo"
analyze	Data pipeline → analysis	"Compute NDVI for California"
visualize	Data pipeline → map	"Show land cover for Denver"
compare	Data pipeline → comparison	"Compare forest cover 2020 vs 2024"
explain	ContextAgent → LLM answer	"What is NDVI?"
monitor	ContextAgent → LLM answer	"Track deforestation trends"

The explain and monitor intents route directly to the ContextAgent, which uses the LLM to provide informative answers.