In the cryptocurrency trading field, analysts with large followings (influencers) produce professional market analysis content daily. Their information streams may seem mysterious, but behind them lies clear data processing logic. Today, we will completely deconstruct this process and teach you how to build your own professional trading analysis system using visual workflow nodes on the Inventor platform.

System Architecture Overview

This workflow is built entirely using visual nodes in the Inventor workflow platform, requiring no complex coding. The core concept is: Data Acquisition → Technical Analysis → Sentiment Analysis → AI Integrated Analysis → Automated Distribution. The entire process can be completed through drag-and-drop nodes and parameter configuration, allowing non-technical personnel to easily build professional trading analysis systems.

Step 1: Building the Data Collection Layer

1. Setting Up the Schedule Trigger Node

Node Type: Schedule Trigger

This node serves as the starting point of the entire system, ensuring the analysis process automatically initiates at fixed time intervals, guaranteeing high frequency and timeliness of the analysis.

1. Building Multi-Timeframe Technical Data Acquisition

The workflow deploys three MarketInfo nodes to obtain data across different time dimensions:

Parameter Explanation:

• period: K-line period - 15 minutes, 1 hour, 1 day
• limit: 200: Retrieves 200 K-lines, sufficient for technical analysis
• $vars.pair: Uses variables to store trading pairs, facilitating easy switching of analysis targets; as an external strategy parameter, it can be set to any asset of interest

1. K-line Data Standardization Node Group

The following three code blocks respectively process K-line data from different timeframes:

Code 1 (15-minute Data Processing):

const result = [];
const data = $input.first().json.result || [];

data.forEach(item => {
  result.push({
    timeframe: "15m",
    candles: item
  });
});

return result;

Code 2 (1-hour Data Processing):

const result = [];
const data = $input.first().json.result || [];

data.forEach(item => {
  result.push({
    timeframe: "1h", 
    candles: item
  });
});

return result;

Code 3 (Daily Data Processing):

const result = [];
const data = $input.first().json.result || [];

data.forEach(item => {
  result.push({
    timeframe: "1d",
    candles: item
  });
});

return result;

Standardization Purpose:

• Adds unique identifiers to data from each timeframe
• Unifies data structure for easier subsequent AI analysis recognition
• Ensures consistency in data flow

1. Data Merge Node

Node Type: Merge Mode: Append mode, sequentially merging data from all three timeframes

This node integrates K-line data from all timeframes into a unified data package, providing a complete technical data foundation for AI analysis.

1. Aggregated K-line Data Processing Node

Node Type: Code Position: Located after the merge node, responsible for integrating multi-timeframe K-line data

Core Code:

const allCandles = [];

for (const item of items){
  allCandles.push(item.json)
}

return [{
  json:{
    allCandles
  }
}];

Code Function Analysis:

• Data Aggregation: Iterates through all data items passed by the merge node, collecting standardized data from each timeframe into the allCandles array
• Structure Unification: Merges K-line data from three timeframes (15-minute, 1-hour, 1-day) into a single data package
• Format Standardization: Outputs a unified JSON structure containing complete K-line information from all timeframes

Step 2: Sentiment Analysis Processing Layer

1. Deploying News Data Acquisition Node

Node Type: HTTP Request

Detailed Configuration:

{
  "parameters": {
    "method": "GET",
    "url": "https://newsapi.org/v2/everything",
    "sendQuery": true,
    "queryParameters": {
      "parameters": [
        {
          "name": "q",
          "value": "Crypto OR Bitcoin OR Coindesk"
        },
        {
          "name": "from",
          "value": "={{ new Date(Date.now() - 3 * 24 * 60 * 60 * 1000).toISOString().split('T')[0] }}"
        }
      ]
    },
    "sendHeaders": true,
    "headerParameters": {
      "parameters": [{
        "name": "x-api-key",
        "value": "xxx"
      }]
    }
  }
}

Key Parameter Analysis:

• q parameter: Search keywords use OR logic to ensure coverage of major cryptocurrency media outlets
• from parameter: Dynamically calculates the date from 3 days ago to ensure retrieval of the latest news
• x-api-key: NewsAPI authentication key, requires registration to obtain

1. News Data Processing Node

Node Type: Code

Core Code:

const articles = $input.first().json.data.articles || [];
const filteredArticles = articles.map(article => ({
  title: article.title,
  description: article.description,
}));
return [{
  json: {
    filteredArticles
  } 
}]

Code Analysis:

• Extracts the articles array returned by NewsAPI
• Retains only title and description fields, removing redundant information like images and links
• Prepares clean text data for subsequent AI analysis

1. Sentiment Analysis AI Node

Node Type: AI Agent LLM Model: Claude model

Core Prompt Configuration:

You are a highly intelligent and precise sentiment analyzer specializing in cryptocurrency markets. You will analyze the sentiment of provided text using a two-part approach:

Short-term Sentiment:
- Assess immediate market reactions, recent news impact, and technical volatility
- Determine sentiment category: "Positive", "Neutral", or "Negative"
- Calculate a numerical score between -1 (extremely negative) and 1 (extremely positive)
- Provide concise reasoning for short-term sentiment (give detailed responses and appropriate headlines for major events and cryptocurrencies)

Long-term Sentiment:
- Evaluate overall market outlook, fundamentals, and regulatory or macroeconomic factors
- Determine sentiment category: "Positive", "Neutral", or "Negative"
- Calculate a numerical score between -1 (extremely negative) and 1 (extremely positive)
- Provide detailed reasoning for long-term sentiment (give detailed responses and appropriate headlines for major events and cryptocurrencies)

Your output must be exactly a JSON object with two keys: "shortTermSentiment" and "longTermSentiment". Each key's value must be an object containing three keys: "category", "score", and "rationale". Do not output any additional text.

Now, analyze the following text and generate your JSON output:
{{ JSON.stringify($json.filteredArticles) }}

Prompt Design Highlights:

• Clearly distinguishes between short-term and long-term sentiment dimensions
• Requires standard JSON output format for easy subsequent processing
• Includes numerical scoring system to quantify sentiment intensity

Step 3: AI Analysis Engine

1. Data Integration Node

Node Type: Code Key Code:

// Initialize containers for each set of data.
const allCandles = [];
let contentData = null;
// Loop over each item from the merge node.
for (const item of items) {
    // If the item has candlestick data, add it to the array.
    if (item.json.allCandles !== undefined) {
        // Assuming item.json.allCandles is an array.
        allCandles.push(...item.json.allCandles);
    }
    // If the item has embedded content (in message.content), store it.
    if (item.json.output !== undefined) {
        contentData = item.json.output;
    }
}
// Return a single item with both candlestick data and content.
return [{
    json: {
        allCandles,
        content: contentData
    }
}];

Integration Logic:

• Merges K-line data with sentiment analysis results
• Prepares complete data package for comprehensive AI analysis
• Ensures data structure meets final analysis requirements

1. Comprehensive Technical Analysis AI Node

Node Type: AI Agent LLM Model: Claude model

Detailed Prompt Configuration:

## Trading Analysis Instructions
**Data Structure:**
{{ $vars.pair}} comprehensive market data:

- Technical Data: {{ JSON.stringify($json["allCandles"]) }}
- Sentiment Analysis: {{ JSON.stringify($json["content"]) }}

K-line Format: Timeframe ("15m", "1h", "1d") + K-line array
Sentiment: Short-term/long-term analysis from cryptocurrency news

**Analysis Framework:**
**Short-term (15m + 1h data):**

- Identify immediate support/resistance levels
- Price action signals + lagging indicators
- Focus on entry/exit timing

**Long-term (1d + 1h data):**

- Primary trend direction
- Structural price levels
- Broader market context

**Output Requirements:**
**Format:** Plain text, Telegram HTML style
**Date:** {{ $vars.pair}} Analysis {{ $now }} (Format: mm/dd/yyyy at xx:xxpm)

**Structure:**
**Spot Trading Recommendations:**
**Short-term:**

- Action: [Buy/Sell/Hold]
- Entry: $X
- Stop Loss: $X
- Target: $X
- Rationale: [2-3 concise sentences covering key signals, indicators, sentiment]

**Long-term:**

- Action: [Buy/Sell/Hold]
- Entry: $X
- Stop Loss: $X
- Target: $X
- Rationale: [2-3 concise sentences covering key signals, indicators, sentiment]

**Leverage Trading Recommendations:**
**Short-term:**

- Position: [Long/Short]
- Leverage: Xx
- Entry: $X
- Stop Loss: $X
- Target: $X
- Rationale: [2-3 concise sentences covering price action, confirmation, sentiment]

**Long-term:**

- Position: [Long/Short]
- Leverage: Xx
- Entry: $X
- Stop Loss: $X
- Target: $X
- Rationale: [2-3 concise sentences covering price action, confirmation, sentiment]

**Key Guiding Principles:**

- Keep each rationale under 50 words
- Focus on actionable insights
- Eliminate redundant explanations
- Prioritize high-confidence signals
- Use direct, concise language

Core Prompt Features:

• Integrates technical data and sentiment analysis
• Multi-timeframe comprehensive analysis
• Provides both spot and leverage trading recommendations
• Outputs HTML format, directly compatible with Telegram
• Emphasizes conciseness and actionability

Step 4: Intelligent Push System

1. Content Splitting Node

Node Type: Code

Splitting Algorithm:

// Get the input text, use an empty string if it doesn't exist.
const inputText = $input.first().json.output || "";

// Validate input type
if (typeof inputText !== "string") {
  throw new Error("Input must be a string");
}

// "Remove '#' and '*' symbols"
const cleanedText = inputText.replace(/[#*]/g, "");

// "Find the position of 'Leverage Trading Recommendations'"
const leveragedIndex = cleanedText.indexOf("Leverage Trading Recommendations");

// "If no split marker is found, split in the original way"
if (leveragedIndex === -1) {
  const mid = Math.ceil(cleanedText.length / 2);
  const firstHalf = cleanedText.substring(0, mid);
  const secondHalf = cleanedText.substring(mid);
  
  return [
    { json: { blockNumber: 1, content: firstHalf } },
    { json: { blockNumber: 2, content: secondHalf } }
  ];
}

// "Split text based on 'Leverage Trading Recommendations'"
const firstBlock = cleanedText.substring(0, leveragedIndex).trim();
const secondBlock = cleanedText.substring(leveragedIndex).trim();

// "Return an array containing two blocks"
return [
  { json: { blockNumber: 1, content: firstBlock } },
  { json: { blockNumber: 2, content: secondBlock } }
];

Splitting Strategy:

• Intelligently identifies “leverage recommendations” keywords for content splitting
• Automatically cleans Markdown symbols to maintain content cleanliness
• Considers Telegram message length limitations
• Maintains logical integrity of analysis reports

1. Telegram Push Node

Node Type: Telegram

Push Features:

• Supports HTML format, maintaining aesthetic formatting of analysis reports
• Automatic block sending to avoid excessively long single messages
• Real-time push notifications ensuring users receive analysis results immediately
• Push to specified groups or channels

This way, we have completed our own dedicated analysis channel, which can provide relevant market analysis and trading recommendation pushes based on the assets we trade.

Core Technical Highlights

1. High-Frequency Data Acquisition

The schedule trigger initiates four data acquisition branches at fixed time intervals, ensuring high timeliness and accuracy of analysis.

2. AI-Driven Professional Analysis

Uses Claude large language model combined with carefully designed prompts to achieve market judgment capabilities comparable to professional analysts.

3. Modular Design

Each function is encapsulated as an independent node, facilitating maintenance, upgrades, and personalized customization.

4. Intelligent Content Processing

Intelligently splits long text based on content structure, ensuring perfect presentation of analysis results across various push channels.

Automated Trading Extensions

If you want to build your own automated trading system, you can add trading execution nodes based on this foundation:

• Trading API Node: Execute corresponding trading operations based on trading recommendations
• Risk Control Node: Set position management and risk control logic
• Order Execution Node: Automatically place orders based on AI analysis results
• Monitoring Node: Real-time tracking of order status and profit/loss conditions

Simply add conditional judgment and trading execution branches after the AI analysis node to achieve a fully automated trading system.

Summary

Through the Inventor workflow visual nodes, we have successfully deconstructed the true nature of crypto influencer information flows. Simply put, those seemingly profound market analyses are backed by such a standardized data processing workflow.

The truth is simple:

• Transparent data sources: News API + market data, nothing mysterious
• Clear analysis logic: AI model + standardized prompts, reproducible and verifiable
• Low technical barrier: Build by dragging nodes, no special skills required
• Extremely low cost: A few API call fees, far lower than paid subscriptions

We don’t need to blindly follow any influencers, nor do we need to pay for so-called “exclusive insider information.” The Inventor quantitative platform has made complex technical analysis accessible, and ordinary people can completely build their own analysis systems.

Most importantly, when you build this system with your own hands, you’ll understand: market analysis is not mysterious, and influencers are not prophets. What they can do, you can do too, and possibly even better.

In this era of technological proliferation, the Inventor platform is making the barrier to investment decision-making lower and lower.

Rather than being someone else’s fan, use the FMZ platform to become the master of your own investment decisions.

Strategy Address: https://www.fmz.com/strategy/509355