Django Quant Tick downloads, and aggregates candlesticks, from high frequency tick data.
Tick data is preferable for analyzing financial markets. Candlesticks aggregated by django-quant-tick are equally informationally dense. Such candles can be useful for analyzing financial markets. As an example, refer to "Low-Frequency Traders in a High-Frequency World: A Survival Guide" and "The Volume Clock: Insights into the High Frequency Paradigm". Lopez de Prado recommends volume candlesticks, however they are are computationally expensive to generate.
Tick data may be downloaded raw, or optionally aggregated. Aggregation can save disk space, and may increase information. There are 3 complementary aggregations. The first is by equal symbol, timestamp, nanoseconds and tick rule. The second is by filtering significant trades, for example at least $1000. The third is by clustering consecutively executed buy or sell orders.
By aggregating, filtering, and clustering raw tick data, volume candlesticks can be computed faster, with little loss in precision.
-
First tick data may be aggregated by equal symbol, timestamp, nanoseconds and tick rule. Aggregating trades in this way can increase information, as they are either orders of size or stop loss cascades. As well, the number of rows can be reduced by 30-50%
-
By filtering aggregated rows, for example only writing a row when an aggregated trade is greater than
significant_trade_filter >= 1000, the number of rows can be reduced more. -
Clustering trades by trade direction, such that a row is created only if the tick rule changes, may further reduce the number of rows.
Whenever possible, data is downloaded from the exchange's AWS S3 repositories. Otherwise, it is downloaded using their REST APIs.
A database, preferably PostgreSQL, is required. Data is saved to the database after aggregation and filtering.
Candles are aggregated at 1 minute intervals, and validated with the exchange's historical candle API.
◻️ Binance REST API (requires API key, which requires KYC)
- Other exchanges validate trade data downloaded from exchanges using candle data provided by exchanges. However, I did not complete KYC, and as a resident of Japan am not currently able to do so. Support is incomplete. Pull requests are welcome.
✅ Bitfinex REST API
✅ BitMEX REST API, and S3 repository
✅ Bybit S3 repository.
- The REST API is no longer paginated, so data may be delayed 24 hours or more.
✅ Coinbase Pro REST API
Note: Exchanges without paginated REST APIs or an S3 repository, will never be supported.
For convenience, django-quant-tick can be installed from PyPI:
pip install django-quant-tick
For deployment, there are Dockerfiles. As well, there are invoke tasks for deployment to Google Cloud Run. Just as easily, the demo could be deployed to a VPS or AWS.
If using GCP, it is recommended to use the Cloud SQL Auth proxy, and run the management commands to collect data from your local machine. Django Quant Tick will upload the trade data to the cloud.
cd demo
invoke start-proxy
python proxy.py trades
Then, configure a Cloud Workflow to collect data in the cloud. There is an example workflow in the invoke tasks.
To use the scripts or deploy to GCP, rename .env.sample to .env, and add the required settings.