Upload README.md with huggingface_hub

README.md

@@ -1,59 +1,59 @@
 ---
 license: mit
-task_categories:
-- time-series-forecasting
-tags:
-- finance
-- market-data
-- order-book
-- high-frequency
-size_categories:
-- 1M<n<10M
+viewer: false
 ---
 
 # LOBSTER Sample Data
 
-This dataset contains sample LOBSTER (Limit Order Book System - The Efficient Reconstructor) market data for several stocks from June 21, 2012.
+Raw LOBSTER (Limit Order Book System) L3 order book data for June 21, 2012.
 
-## Dataset Description
+**Important:** These are raw CSV files without headers. They cannot be automatically loaded by HuggingFace viewers. Use the loading code below with custom schemas.
 
-LOBSTER provides limit order book data reconstructed from NASDAQ TotalView-ITCH messages. This sample dataset includes:
+## Available Datasets
 
-- **AAPL** (Apple Inc.) - 50 levels
-- **MSFT** (Microsoft Corporation) - 50 levels
-- **SPY** (SPDR S&P 500 ETF Trust) - 50 levels
-- **GOOG** (Google Inc.) - 10 levels
+- **AAPL**: 1, 5, 10, 30, 50 levels
+- **AMZN**: 1, 5, 10 levels  
+- **GOOG**: 1, 5, 10 levels
+- **INTC**: 1, 5, 10 levels
+- **MSFT**: 1, 5, 10, 30, 50 levels
+- **SPY**: 30, 50 levels
 
 ## File Structure
 
-Each ticker has two CSV files:
-
-1. **Message file** (`*_message_*.csv`): Contains all order book events
+1. **Message file** (`*_message_*.csv`): Order book events (6 columns, no header)
    - Columns: time, event_type, order_id, size, price, direction
-   - Event types: 1=New Order, 2=Cancellation, 3=Deletion, 4=Execution(Visible), 5=Execution(Hidden), 7=Trading Halt
+   - Types: 1=New, 2=Cancel, 3=Delete, 4=Exec(Visible), 5=Exec(Hidden), 7=Halt
 
-2. **Orderbook file** (`*_orderbook_*.csv`): Contains order book snapshots after each event
+2. **Orderbook file** (`*_orderbook_*.csv`): Order book snapshots (4N columns, no header)
    - Columns: ask_price_1, ask_size_1, bid_price_1, bid_size_1, ... (up to N levels)
-   - Prices are in fixed-point format (multiply by 10^-4 for actual price)
+   - Prices in fixed-point (multiply by 10^-4)
 
 ## Usage
 
 ```python
 from huggingface_hub import hf_hub_download
-import pandas as pd
+import polars as pl
 
 message_file = hf_hub_download(
-    repo_id="YOUR_USERNAME/lobster-data",
+    repo_id="totalorganfailure/lobster-data",
     filename="LOBSTER_SampleFile_AAPL_2012-06-21_50/AAPL_2012-06-21_34200000_37800000_message_50.csv",
     repo_type="dataset"
 )
 
-orderbook_file = hf_hub_download(
-    repo_id="YOUR_USERNAME/lobster-data",
-    filename="LOBSTER_SampleFile_AAPL_2012-06-21_50/AAPL_2012-06-21_34200000_37800000_orderbook_50.csv",
-    repo_type="dataset"
-)
-
-messages = pd.read_csv(message_file, header=None)
-orderbook = pd.read_csv(orderbook_file, header=None)
+# Define schema for message file
+MESSAGE_SCHEMA = {
+    "time": pl.Float64,
+    "type": pl.Int8,
+    "order_id": pl.Int64,
+    "size": pl.Int32,
+    "price": pl.Int64,
+    "direction": pl.Int8,
+}
+
+messages = pl.read_csv(message_file, has_header=False, schema=MESSAGE_SCHEMA)
 ```
+
+## More Information
+
+- https://lobsterdata.com/
+- https://lobsterdata.com/info/DataSamples.php