PlatCheck - Warframe Item Sell Price Estimator

A simple web app that estimates the optimal price to sell warframe items at using Ridge Regression.

Features

• Price Trends: View 1-day, 7-day, and 30-day median prices
• Fair Price: ML-based price estimation using Ridge Regression
• Search: Autocomplete search for Warframe items

Tech Stack

• Frontend: Next.js with TypeScript and Tailwind CSS
• Backend: Next.js API routes
• ML: Python + scikit-learn (Ridge Regression)
• Data Source: Warframe.market API v2

Setup

Install Dependencies

# Install dependencies
npm install

# Create Python virtual environment
python -m venv venv

# Activate virtual environment
source venv/bin/activate

# Install Python dependencies
pip install "numpy>=1.24.0" "scikit-learn>=1.3.0" "requests>=2.31.0"

Train the Model

# Make sure virtual environment is activated
source venv/bin/activate

# Run training script
python scripts/train_model.py

This will:

• Fetch historical price data from the Warframe.market API
• Train a Ridge Regression model
• Save the model to model/model.json

Note: The API has rate limiting, so this may take a while.

3. Run the dev server

npm run dev

Open http://localhost:3000.

Why Ridge Regression?

Ridge Regression is a technique that finds patterns in historical prices and uses them to estimate what an item should cost.

The model uses three pieces of information to make predictions:

• median_1d: The median price from the last 1 day (current market)
• median_7d: The median price from the last 7 days (short-term trends)
• median_30d: The median price from the last 30 days (longer-term trends)

These medians are weighted so that days with more trading volume are more important in the calcuations.

The optimal sell price is then calculated with this formula:

Sell Price = Base Price + (Weight1 * median_1d) + (Weight2 * median_7d) + (Weight3 * median_30d)

The weights are learned through training.

Training Workflow

1. Get Real-Time Data: The training script downloads historical prices for up to 200 items from the Warframe.market API
2. Clean the Data: The script:
   • Removes price outliers
   • Ignores days with low to no trading volume
3. Build Patterns: The script looks at examples of previous trades and maps:
   • Inputs: 1-day, 7-day, and 30-day medians
   • Target: the next day’s median price
4. Train Model:
   • Ridge Regression learns the best weights to explain how past prices relate to future ones
5. Save the Model:
   • The weights are stored in a JSON file

Prediction Workflow

When a user searches for an item:

1. Get recent prices from Warframe.market API
2. Compute 1/7/30 day trends
3. Use the trained model on the trends to compute the optimal sell price
4. Bounds-check the value to within current buy/sell values
5. Return the result