Make Mixpanel Data

🤖 What is this?

Generate fake Mixpanel data quickly and easily with a simple CLI in Node.js. Create events, user profiles, groups profiles, SCD data, mirror tables, and lookup tables; basically everything you need for testing + demoing mixpanel analysis features.

the CLI looks like this:

under the hood, make-mp-data is modeling data adherent to match Mixpanel's data model, giving you the tools you need for robust, realistic field ready test data.

🚀 Quick Start

Basic Usage

Generate events and users, and write them to CSV files:

npx make-mp-data 

Customizing Output

Generate 10,000 events and 100 users over the last 90 days, and write the output as JSON:

npx make-mp-data --numUsers 100 --numEvents 10000 --numDays 90 --format json

Complex Data Models

Create a comprehensive dataset including events, users, groups, SCD, and lookup tables:

npx make-mp-data --complex

Send Data to Mixpanel

Generate and send data directly to Mixpanel using your project token:

npx make-mp-data --token 1234

Help and Options

Need more info on available options?

npx make-mp-data --help

🔧 Schema

To choose your own event and property names and values, specify a custom data model and additional options:

npx make-mp-data [dataModel.js] [options]

Example:

npx make-mp-data myDataSpec.js --token 1234 --numDays 30 --numUsers 1000 --numEvents 1000000

where myDataSpec.js exports a JS object of this shape ... (see ./schemas for examples)

Data Models

Check out ./schemas/ for example dataModel.js files to get started quickly.

🛠️ CLI Options

Here's a breakdown of the CLI options you can use with make-mp-data:

• --numUsers: Number of users to generate.
• --numEvents: Number of events to generate.
• --numDays: Number of days over which to spread the generated data.
• --format: Output format (csv or json).
• --token: Mixpanel project token for direct data import.
• --region: Mixpanel data region (US, EU).
• --writeToDisk: Whether to write the data to disk (true or false).
• --verbose: Enable verbose logging.
• --complex: create a complex set models including groups, SCD, and lookup tables.
• --simple: create a simple dataset including events, and users

⏱️ TimeSoup — Realistic Time Distributions

TimeSoup controls how events are distributed across time. Out of the box, it produces realistic day-of-week and hour-of-day patterns derived from real Mixpanel data (weekday-heavy, Saturday valley, morning peak).

Presets

Use a preset string for quick configuration:

const config = {
  soup: "growth",    // default — gradual uptrend with weekly cycle
  // ...
};

| Preset | Pattern | Use Case | |--------|---------|----------| | "steady" | Flat, minimal variation | Enterprise B2B, utility apps | | "growth" | Gradual uptrend + weekly cycle | General purpose (default) | | "spiky" | Dramatic peaks and valleys | Gaming, social media, viral products | | "seasonal" | 3-4 major waves | E-commerce, education | | "global" | Flat DOW + flat HOD | Global SaaS, infrastructure tools | | "churny" | Flat, no growth trend | Declining products (pair with churn hooks) | | "chaotic" | Wild variation | Anomaly detection, incident response |

Custom Configuration

Override specific parameters or use a preset as a base:

// Preset + overrides
soup: { preset: "spiky", deviation: 5 }

// Fully custom
soup: {
  peaks: 200,
  deviation: 2,
  mean: 0,
  dayOfWeekWeights: [0.637, 1.0, 0.999, 0.998, 0.966, 0.802, 0.528],  // [Sun..Sat]
  hourOfDayWeights: [/* 24 values, index 0 = midnight UTC */],
}

// Disable cyclical patterns entirely
soup: { dayOfWeekWeights: null, hourOfDayWeights: null }

🎯 Hooks — Engineering Trends in Data

Hooks let you engineer deliberate, discoverable patterns into your generated data — things like "premium users convert 2x better" or "there was a service outage during days 40-47." A hook is a single transform function on your dungeon config:

const config = {
  // ...events, funnels, etc.
  hook: function(record, type, meta) {
    // Modify events based on type
    if (type === "event" && record.event === "purchase") {
      record.amount = record.amount * 2; // boost purchase amounts
    }
    
    // Use "everything" to correlate across a user's full event stream
    if (type === "everything") {
      const hasPurchase = record.some(e => e.event === "purchase");
      for (const event of record) {
        event.is_buyer = hasPurchase; // tag all events for this user
      }
      return record;
    }
    
    return record;
  }
};

Hook types: event, user, everything, funnel-pre, funnel-post, scd-pre, ad-spend, group, mirror, lookup.

See dungeons/harness/ for comprehensive examples with 8 hooks each across gaming, fintech, education, food delivery, media, social, and SaaS verticals.

📄 Examples

Check out the examples directory for sample data models:

ls ./schemas/

These models provide a great starting point for creating your own custom data generation scripts.

🤝 Contributing

Feel free to fork this repository and submit pull requests. Contributions are always welcome!

For any issues or feature requests, please create an issue on the GitHub repository.