Collect Free Data for US Stocks¶

Free data for US stocks is available through two distinct, complementary datasets. The free minute bundle provides minute and daily price data from 2007 to the present for a small selection of stocks. The "learning bundle" provides daily price history from 2007-2011 for all US stocks (including stocks that later delisted from the exchange). The free minute bundle provides a large date range for a small number of stocks, while the learning bundle provides a smaller date range for a large number of stocks. Both datasets are utilized in the lecture series.

The datasets are stored as Zipline bundles, a type of local database. (Zipline is a backtesting library that also provides a database technology for storing and querying price data. See the usage guide for more information.)

Free minute bundle¶

Start by creating an empty bundle called 'usstock-free-1min', specifying free=True:

In [1]:

from quantrocket.zipline import create_usstock_bundle
create_usstock_bundle("usstock-free-1min", free=True)

Out[1]:

{'status': 'success', 'msg': 'successfully created usstock-free-1min bundle'}

Then collect the data, which in Zipline bundle terminology is referred to as ingestion:

In [2]:

from quantrocket.zipline import ingest_bundle
ingest_bundle("usstock-free-1min")

Out[2]:

{'status': 'the data will be ingested asynchronously'}

This runs in the background. Open flightlog to monitor the progress. You can open flightlog from the JupyterLab launcher or by executing the following IPython magic command:

The following messages will indicate completion:

quantrocket.zipline: INFO [usstock-free-1min] Ingesting minute bars for 9 securities in usstock-free-1min bundle
quantrocket.zipline: INFO [usstock-free-1min] Ingesting daily bars for usstock-free-1min bundle
quantrocket.zipline: INFO [usstock-free-1min] Ingesting adjustments for usstock-free-1min bundle
quantrocket.zipline: INFO [usstock-free-1min] Ingesting assets for usstock-free-1min bundle
quantrocket.zipline: INFO [usstock-free-1min] Completed ingesting data for 9 securities in usstock-free-1min bundle

Learning bundle¶

Collecting the learning dataset is similar. Create a bundle called 'usstock-learn-1d', specifying learn=True:

In [4]:

from quantrocket.zipline import create_usstock_bundle
create_usstock_bundle("usstock-learn-1d", learn=True)

Out[4]:

{'status': 'success', 'msg': 'successfully created usstock-learn-1d bundle'}

Then ingest the data:

In [5]:

from quantrocket.zipline import ingest_bundle
ingest_bundle("usstock-learn-1d")

Out[5]:

{'status': 'the data will be ingested asynchronously'}

This dataset only takes a few seconds to ingest:

quantrocket.zipline: INFO [usstock-learn-1d] Completed ingesting daily history for all stocks from 2007-2011