Pandas ML Quant Data Provider

An independent module used to fetch time series and other data used for quant finance studies.

Installation

Data can always be fetched instant without any local data base. However if you need to screen across assets for matching criteria then a local database is needed. Developing models often depend on some premise. Like for a crash detection model you would want to use instruments which indeed suffered a crash. Therefore some data is stored into a data base using dolthub. Querying this data requires you to install dolt by following their installation instructions.

pip install pandas-quant-data-provider
 

!! NOTE !! We need to keep data libraries like yfincance up do date alsmost every minute. Whenever there is a change in the api the library has to react. This happens more often as we whish, especially with yahoo finance. This is why this library gets pip installed -U on every import. This means for server based solutions on every restart. Make sure you use all pandas-ml* libraries inside a virtual env.

Example:

# monkey patch pandas 
from pandas_quant_data_provider import pd, YAHOO, INVESTING, CRYPTO_COMPARE

# fetch data from various data sources 
#   * fetches all available dates
#   * caches data for 10 minutes
df = pd.fetch_timeseries({
    YAHOO: ["SPY", "DIA"],
    INVESTING: ["index::NYSE Tick Index::united states", "bond::U.S. 30Y::united states"],
    CRYPTO_COMPARE: ["BTC"]
})

df.tail()

PS If you are not familiar with pandas MultiIndex, you can watch this video: How do I use the MultiIndex in pandas?

Installation

Follow the instructions on https://github.com/KIC/pandas-ml-quant

Documentation

Check out the notebooks at https://github.com/KIC/pandas-ml-quant/blob/master/notebooks