Code Examples

Predict sales based on marketing spend with automatic feature selection and diagnostics.

# Use MCP Analytics to run linear regression
mcp.tools_run(
  tool_name="linear_regression",
  taskList={
    "inputs": {
      "dataset": "https://your-data-url.com/sales.csv",
      "userContext": {
        "objective": "Predict sales based on marketing spend",
        "company": "Your Company"
      },
      "target": "sales_amount",
      "features": ["tv_spend", "digital_spend", "print_spend"]
    }
  },
  generate_insights=true
)

Upload and encrypt your local datasets for analysis.

# Step 1: Generate upload token
token_info = mcp.datasets.upload(
  expires_in=1800,  # 30 minutes
  metadata={
    "description": "Q4 2024 Sales Data",
    "type": "sales"
  }
)

# Step 2: Use the provided curl command to upload
# Step 3: Analyze your encrypted dataset
mcp.tools_run(
  tool_name="data_profiling",
  taskList={
    "inputs": {
      "dataset": f"uuid://{token_info.uuid}:{token_info.key}"
    }
  }
)

Forecast future values with seasonal patterns and confidence intervals.

mcp.tools_run(
  tool_name="arima_forecast",
  taskList={
    "inputs": {
      "dataset": "monthly_revenue.csv",
      "userContext": {
        "objective": "Forecast next 6 months revenue with seasonality"
      },
      "target": "revenue",
      "date_column": "month",
      "periods": 6,
      "seasonal": true,
      "confidence": 0.95
    }
  }
)

Find insights across all your analyses using natural language.

# Search for related analyses
results = mcp.reports.search(
  semantic_query="customer churn prediction models with high accuracy",
  threshold=0.8,
  date_from="2024-01-01",
  limit=5
)

# Get full report details
report = mcp.reports.search(
  job_ids=[results[0].id],
  include_data=true,
  keys=["model_metrics", "feature_importance"]
)

# Generate shareable link
share_url = mcp.reports.view(
  processing_id=results[0].id,
  expires_in=86400  # 24 hours
)