Exploring optimal order execution policies on limit order books
Execute the script scripts/scrapers/download_s3_data.py
with command line arguments like:
python download_s3_data.py --pair=XBTZUSD --bucket=my-bucket --save_directory=/path/to/save/location --access_key_id=KEY_ID --secret_access_key=SECRET_KEY
Valid parameters for pair argument: {ALL, USDTZUSD, XETHZUSD, XXMRZUSD, XXRPZUSD, XREPZUSD, XBTZUSD}
The bucket name is the S3 bucket containing our data set. A set of AWS credentials is required to access it. Valid parameters for the bucket argument are: {limit-order-book, recent-trades}
Check scripts/utils/data_loading.py
for the load_data(...)
method. You can specify the directory containing the parquet data and the method will combine the separate files into a single data frame. In case of the limit order book data, the whole data set might not fit into memory. In that case, datetime parameters can be passed to only select the subset of data scraped within those dates. Exact instructions are in the docstring.