GitHub

Query Loops

Query loops are a powerful feature of GeoRetina AI that allows you to build on previous query results, enabling multi-step analysis and progressive refinement of your insights.

What are Query Loops?

Query loops let you use the results of one query as the input for subsequent queries. This creates an iterative analysis workflow where you can:

  • Narrow down from broad initial results to specific insights
  • Apply multiple filters or conditions in sequence
  • Build complex analytical workflows through simple conversational steps
  • Refine your analysis based on intermediate findings

How Query Loops Work

When you run an initial query, GRAI returns a set of results. Rather than starting a completely new query, you can refer to "these results" or "the selected features" in your follow-up questions, creating a loop of increasingly refined analysis.

Example Workflow

  1. Initial query: "Show me all schools in this region"

    Initial query example

  2. View initial results: GRAI returns matching features

    Initial query results

  3. Follow-up query: "Of these schools, show me only those with more than 500 students"

  4. Further refinement: "Among these results, which ones are within 1km of a park?"

  5. Final analysis: "For these remaining schools, what's the average distance to the nearest hospital?"

    Final query analysis results

Each step builds on the previous result set, creating a logical chain of analysis.

Benefits of Query Loops

Query loops offer several advantages:

  • Simplified complex analysis: Break down complex questions into manageable steps
  • Interactive exploration: Adjust your approach based on intermediate findings
  • Reduced data processing: Focus only on relevant subsets of data
  • Intuitive workflow: Mirrors how people naturally think about multi-step analysis
  • Transparent process: Each step is clearly documented in the conversation

Common Query Loop Patterns

Filter Refinement

Start with a broad set and progressively apply filters:

"Show me all buildings in this area"
"Of these buildings, show only commercial properties"
"Among these commercial properties, which ones were built after 2000?"
"Of these newer buildings, which ones have solar panels installed?"

Comparative Analysis

Compare different subsets of data:

"Show me all parks in the city"
"Now show me parks that have playgrounds"
"Compare the average size of parks with playgrounds to those without"

Spatial Relationships

Explore how features relate to each other:

"Show me all hospitals in the region"
"Now show me residential areas within 2km of these hospitals"
"Of these residential areas, which ones have the highest population density?"

Sequential Calculations

Perform calculations at each step:

"Show me all agricultural parcels in the county"
"Calculate the total area of these parcels"
"Of these parcels, show only those growing corn"
"What percentage of the total agricultural area is used for corn?"

Tips for Effective Query Loops

To get the most out of query loops:

  • Start broad, then narrow: Begin with a larger dataset and refine
  • Use clear references: Phrases like "these results," "these features," or "among these" help GRAI understand you're building on previous results
  • Verify intermediate steps: Make sure each step gives you the expected results before proceeding
  • Combine with other data sources: Bring in additional datasets at different stages of your analysis
  • Save important results: Export critical findings at various points in your analysis workflow

Limitations

Some current limitations to be aware of:

  • Query loops work best with vector data (as opposed to raster/satellite data)
  • Very large result sets may need to be further filtered before detailed analysis
  • Complex chains with many steps might require occasional clarification
  • Some advanced spatial operations may not be available for all data types

Examples of Real-World Applications

Urban Planning

"Show me all vacant lots in the downtown area"
"Of these lots, which ones are zoned for commercial use?"
"Among these commercial lots, which ones are within 500m of public transportation?"
"For these well-connected lots, what's the current market value per square meter?"

Environmental Analysis

"Show me all wetlands in the county"
"Of these wetlands, which ones have decreased in size over the past 10 years?"
"For these shrinking wetlands, are there any industrial facilities within 1km?"
"Among these facilities, which ones have reported chemical discharges?"

Infrastructure Planning

"Show me all neighborhoods in the city"
"Of these neighborhoods, which ones have the lowest coverage of fiber internet?"
"Among these underserved areas, what's the population density?"
"In these high-density, underserved areas, where would be the optimal locations for new fiber nodes?"

Getting Started with Query Loops

The best way to master query loops is through practice. Start with simple two-step analyses and gradually build up to more complex workflows as you become familiar with the capabilities.