> ## Documentation Index > Fetch the complete documentation index at: https://docs.markifact.com/llms.txt > Use this file to discover all available pages before exploring further. # Analyze Data > Perform data analysis with AI insights, generate charts & graphs, aggregate data, calculate custom metrics Perform data analysis with AI insights, generate charts & graphs, aggregate data, calculate custom metrics. You need to provide dataset from other nodes and a prompt, it can return either a markdown response (graphs, tables, etc.) or structured data you provide in schema, structure data can be used in nodes like Google Sheets, Google Slides Table, etc. Note: charts will be returned as image URLs in markdown — no other formats are supported. | | | | ------------------- | ----------------- | | **App** | AI | | **Operation ID** | `ai_analyze_data` | | **Type** | Action | | **Connection** | None | | **Credits per run** | 2 | | **Versions** | 1, 2 (default: 2) | ## Inputs | Field | Type | Required | Default | Description | | --------------- | -------------------------- | -------- | ------- | --------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------- | | `prompt` | string | Yes | - | Instructions for analyzing the data. This is a multi-line text area where you can: - Describe what analysis to perform - Ask specific questions about the data - Request insights, trends, or patterns - Use line breaks to structure complex analysis requests - Request charts and visualizations if needed. IMPORTANT RULES: - NEVER include actual data, datasets, or data content in this prompt field - ALWAYS use the separate 'data' field below for all datasets and data content - This prompt should ONLY contain analysis instructions, questions, and what insights you want - Even if the user provides data in their request, extract it and put it in the 'data' field instead - Charts will be returned as image URLs in markdown — no other formats are supported. - Do NOT request technical details for charts like image resolution or format — just describe the chart you want - When using 'schema\_fields', do not ask the AI to return JSON, it will be returned automatically by default. | | `model` | string or SelectableOption | Yes | - | The AI model to use for data analysis. Some available options: 'gpt-5.2' - Default, best for data analysis 'gpt-4.1' - Support larger context windows, up to 1m tokens | | `data` | string | Yes | - | The data to analyze. This is mainly the output from upstream workflow nodes containing datasets, tables, or structured information. Reference data from other nodes using \{\{nodeId\_\_data}} syntax. Example: \{\{googleSheetsNodeId\_\_data}} For multiple datasets, add the dataset name before, for example: Meta Ads: \{\{metaAdsNodeId\_\_data}} Google Analytics: \{\{googleAnalyticsNodeId\_\_data}} | | `schema_fields` | array of SchemaField | No | - | - Defines a LIST of objects, where each object represents a field definition for the structured output. You are explicitly defining the keys of each object in the output schema. - Each object in the list must include: - name: The key name that will appear in the structured output (e.g., "clicks", "impressions"). - type: Must be one of: `text`, `number`, `boolean`, `list`. Note that `list` represents a list of strings. - description – A clear explanation of what this field represents. - The schema fields can be referenced by downstream nodes using: \{\{nodeId\_\_data.field}}. - Use this when you need the analysis results in a consistent format for downstream nodes like Google Sheets, Slides, etc. - Don't use it to generate summary/markdown tables, this should be normal text output using the `prompt` field. | | `force_python` | boolean | No | - | Force the AI to use Python code for data analysis instead of just text analysis. When enabled, the AI will write and execute Python code to: - Perform statistical calculations - Generate charts and visualizations - Handle complex data manipulations - Provide more precise numerical analysis | ### SchemaField | Field | Type | Required | Default | Description | | ------------- | ------ | -------- | ------- | ----------------------------------------------------------------------------------- | | `name` | string | Yes | - | The name of the field. This becomes the key in the output object. | | `type` | string | Yes | - | Allowed values: 'text', 'number', 'boolean', 'list'. The list is a list of strings. | | `description` | string | Yes | - | A clear explanation of what this field represents. | ### SelectableOption | Field | Type | Required | Default | Description | | ------- | ------ | -------- | ------- | ----------- | | `value` | string | Yes | - | | | `label` | string | Yes | - | | ## Output **Type**: `str | List[Dict]` Returns markdown text with analysis/charts. If schema\_fields provided, it returns structured data matching the schema in list\[Dict] format. **Fields**: dynamic (depend on the inputs) **Example**: ```json theme={"dark"} [ { "metric": "sessions", "value": 1500 } ] ```