Aller au contenu principal

Insights

Category: AI & ML
Node id: insights

AI-powered data analysis and insights generation with trend and anomaly detection

What it does​

  • Comprehensive data analysis with trend and anomaly detection
  • Confidence-scored insights with severity classification
  • Configurable analysis types (comprehensive, trends, anomalies, recommendations)
  • Historical data and benchmark comparison
  • Visualization suggestions for key insights
  • Structured JSON output with evidence-based findings

Use cases​

  • Analyzing patient population health trends
  • Detecting anomalies in clinical or operational data
  • Generating actionable recommendations from healthcare metrics
  • Comparing performance against industry benchmarks
  • Executive reporting with data-driven insights

Runtime​

  • Performance tier: variable

Input ports​

NameTypeRequiredDescription
dataanyYesData to analyze (string or structured)
contextanyNoAdditional context about the data
objectivesanyNoSpecific analysis objectives
metricsanyNoKey metrics to focus on
historical_dataanyNoHistorical data for comparison
benchmarkanyNoBenchmark data for comparison
constraintsanyNoAnalysis constraints or limitations

Output ports​

NameTypeRequiredDescription
resultjsonNoMain result containing insights, summary, recommendations, and visualizations
metadatajsonNoExecution metadata including tokens, insight_count, and severity counts
successbooleanNoWhether the operation succeeded

Settings​

Basic​

SettingNameTypeRequiredDefaultAllowed valuesDescription
ProviderproviderenumNoopenaiopenai, anthropicLLM provider
ModelmodelstringNo""Model name (auto-selected if empty)
Analysis Typeanalysis_typeenumNocomprehensivecomprehensive, trends, anomalies, recommendationsType of analysis to perform
TimeframetimeframestringNo""Timeframe for analysis
Compare Withcompare_withstringNo""Comparison method (baseline, previous_period, industry_average)
Include Visualizationinclude_visualizationbooleanNofalseSuggest visualizations for insights

Auth​

SettingNameTypeRequiredDefaultAllowed valuesDescription
API Keyapi_keysecretNoAPI key for the LLM provider
Base URLbase_urlstringNo""Custom API base URL

Advanced​

SettingNameTypeRequiredDefaultAllowed valuesDescription
TemperaturetemperaturenumberNo0.3min 0; max 1Sampling temperature
Max Tokensmax_tokensnumberNo4000min 1Maximum tokens for response

Errors​

CodeDescription
DATA_REQUIREDNo data input was provided for analysis
LLM_REQUEST_FAILEDLLM API call failed during analysis
UNSUPPORTED_PROVIDERProvider must be openai or anthropic

Runtime notes​

  • Each insight includes a confidence score (0-1); filter on confidence >= 0.8 for high-confidence results
  • LLM response is parsed as JSON — if parsing fails, a single fallback insight wraps the raw text
  • metadata.tokens tracks LLM token usage for cost monitoring
  • visualizations array is only populated when include_visualization config is true

See Build any workflow for how to put nodes together and Ports, types and data flow for how port types connect.

Quelque chose n'est pas clair ? Envoyez vos commentaires sur la documentation par courriel.