The Macrosynergy package supports financial market research and the development of trading strategies based on formats and conventions of the J.P. Morgan Macrosynergy Quantamental System (JPMaQS). JPMaQS provides quantitative-fundamental (quantamental) and market data in simple daily formats in accordance with the information state of markets. The Macrosynergy package consists of six sub-packages:
- management: simulates, analyses and reshapes standard quantamental dataframes.
- panel: analyses and visualizes panels of quantamental data.
- signal: transforms quantamental indicators into trading signals and does naive analysis.
- pnl: constructs portfolios based on signals, applies risk management and analyses realistic PnLs.
- download: interface for downloading data from JP Morgan DataQuery, with main module jpmaqs.py.
- dataquery: [DEPRECATED] interface for downloading data from JP Morgan DataQuery, with main module api.py.
The easiest method for installing the package is to use the PyPI installation method:
pip install macrosynergyAlternatively for the cutting edge development version, install the package from the develop branch as
pip install git+https://github.com/macrosynergy/macrosynergy@developTo download data from JP Morgan DataQuery, you can use the JPMaQSDownload Object together with your OAuth authentication credentials (default):
import pandas as pd
from macrosynergy.download import JPMaQSDownload
with JPMaQSDownload(
client_id="<dq_client_id>",
client_secret="<dq_client_secret>"
) as downloader:
data = downloader.download(tickers="EUR_FXXR_NSA",
start_date="2022-01-01")
assert isinstance(data, pd.DataFrame) and not data.empty
assert data.shape[0] > 0
data.info()Alternatively, you can also the certificate and private key pair, to access DataQuery as:
import pandas as pd
from macrosynergy.download import JPMaQSDownload
with JPMaQSDownload(
oauth=False,
username="<dq_username>",
password="<dq_password>",
crt="<path_to_dq_certificate>",
key="<path_to_dq_key>"
) as downloader:
data = downloader.download(tickers="EUR_FXXR_NSA",
start_date="2022-01-01")
assert isinstance(data, pd.DataFrame) and not data.empty
assert data.shape[0] > 0
data.info()Both of the above example will download a snippet of example data from the premium JPMaQS dataset of the daily timeseries of EUR FX excess returns.
Using the API you can also access a panel of tickers from different countries like so.
import pandas as pd
from macrosynergy.download import JPMaQSDownload
cids = ['EUR','GBP','USD']
xcats = ['FXXR_NSA','EQXR_NSA']
tickers = [cid+"_"+xcat for cid in cids for xcat in xcats]
with JPMaQSDownload(
client_id="<dq_client_id>",
client_secret="<dq_client_secret>"
) as downloader:
data = downloader.download(tickers=tickers,
start_date="2022-01-01")
assert isinstance(data, pd.DataFrame) and not data.empty
assert data.shape[0] > 0
data.info()Since a lot of institutions use a proxy server to connect to the internet; the JPMaQSDownload object can be configured to use a proxy server.
It is also possible to use a proxy server with the Dataquery interface. Here's an example:
import pandas as pd
from macrosynergy.download import JPMaQSDownload
cids = ['EUR','GBP','USD']
xcats = ['FXXR_NSA','EQXR_NSA']
tickers = [cid+"_"+xcat for cid in cids for xcat in xcats]
oauth_proxy="https://secureproxy.example.com:port"
proxy = {"https": oauth_proxy}
# or proxy = {"http": "http://proxy.example.com:port"}
with JPMaQSDownload(
client_id = "<dq_client_id>",
client_secret = "<dq_client_secret>",
proxy = proxy
) as downloader:
data = downloader.download(tickers = tickers, start_date="2022-01-01")
assert isinstance(data, pd.DataFrame) and not df.emptyor,
...
proxies = {
"http": "http://proxy.example.com:port",
"https": "https://secucreproxy.example.com:port",
}
with JPMaQSDownload(
client_id = "<dq_client_id>",
client_secret = "<dq_client_secret>",
proxy = proxies
) as downloader:
data = downloader.download(tickers = tickers)
...In order to use the rest of the package without access to the API you can simulate quantamental data using the management sub-package.
from macrosynergy.management.simulate_quantamental_data import make_qdf
cids = ['AUD', 'GBP', 'NZD', 'USD']
xcats = ['FXXR_NSA', 'FXCRY_NSA', 'FXCRR_NSA', 'EQXR_NSA', 'EQCRY_NSA', 'EQCRR_NSA',
'FXWBASE_NSA', 'EQWBASE_NSA']
df_cids = pd.DataFrame(index=cids, columns=['earliest', 'latest', 'mean_add',
'sd_mult'])
df_cids.loc['AUD'] = ['2000-01-01', '2022-03-14', 0, 1]
df_cids.loc['GBP'] = ['2001-01-01', '2022-03-14', 0, 2]
df_cids.loc['NZD'] = ['2002-01-01', '2022-03-14', 0, 3]
df_cids.loc['USD'] = ['2000-01-01', '2022-03-14', 0, 4]
df_xcats = pd.DataFrame(index=xcats, columns=['earliest', 'latest', 'mean_add',
'sd_mult', 'ar_coef', 'back_coef'])
df_xcats.loc['FXXR_NSA'] = ['2010-01-01', '2022-03-14', 0, 1, 0, 0.2]
df_xcats.loc['FXCRY_NSA'] = ['2010-01-01', '2022-03-14', 1, 1, 0.9, 0.2]
df_xcats.loc['FXCRR_NSA'] = ['2010-01-01', '2022-03-14', 0.5, 0.8, 0.9, 0.2]
df_xcats.loc['EQXR_NSA'] = ['2010-01-01', '2022-03-14', 0.5, 2, 0, 0.2]
df_xcats.loc['EQCRY_NSA'] = ['2010-01-01', '2022-03-14', 2, 1.5, 0.9, 0.5]
df_xcats.loc['EQCRR_NSA'] = ['2010-01-01', '2022-03-14', 1.5, 1.5, 0.9, 0.5]
df_xcats.loc['FXWBASE_NSA'] = ['2010-01-01', '2022-02-01', 1, 1.5, 0.8, 0.5]
df_xcats.loc['EQWBASE_NSA'] = ['2010-01-01', '2022-02-01', 1, 1.5, 0.9, 0.5]
data = make_qdf(df_cids, df_xcats, back_ar=0.75)The management sub-package can also be used to check which data is available in the dataframe.
from macrosynergy.management.check_availability import check_availability
filt_na = (data['cid'] == 'USD') & (data['real_date'] < '2015-01-01')
data_filt.loc[filt_na, 'value'] = np.nan
check_availability(df=data_filt, xcats=xcats, cids=cids)You can also use the built-in function to reshape the data depending on the dates or tickers of your choice.
data_reduced = reduce_df(data, xcats=xcats[:-1], cids=cids[0],
start='2012-01-01', end='2018-01-31')The basket class is used to calculate the returns and carries of financial contracts using various methods, a basket is created as so.
from macrosynergy.panel.basket import Basket
black = {'AUD': ['2010-01-01', '2013-12-31'], 'GBP': ['2010-01-01', '2013-12-31']}
contracts = ['AUD_FX', 'AUD_EQ', 'NZD_FX', 'GBP_EQ', 'USD_EQ']
gdp_figures = [17.0, 17.0, 41.0, 9.0, 250.0]
basket_1 = Basket(
df=data, contracts=contracts_1, ret="XR_NSA", cry=["CRY_NSA", "CRR_NSA"],
blacklist=black
)
basket_1.make_basket(weight_meth="equal", max_weight=0.55, basket_name="GLB_EQUAL")Using the basket class you have access to the methods such as visulasing the weights associated with each contract, or returning the weight or basket.
basket_1.return_basket()
basket_1.return_weights()
basket_1.weight_visualiser(basket_name="GLB_EQUAL")You can also calculate and visualise the following and more with built-in functions.
from macrosynergy.panel.historic_vol import historic_vol
data_historic = historic_vol(
data, cids=cids, xcat='FXXR_NSA', lback_periods=21, lback_meth='ma', half_life=11,
remove_zeros=True)from macrosynergy.panel.make_zn_scores import make_zn_scores
z_mean = make_zn_scores(data, xcat='FXXR_NSA', sequential=True, cids=cids,
blacklist=black, iis=False, neutral='mean',
pan_weight=0.5, min_obs=261, est_freq="w")
z_median = make_zn_scores(data, xcat='FXXR_NSA', sequential=True, cids=cids,
blacklist=black, iis=False, neutral='median',
pan_weight=0.5, min_obs=261, est_freq="d")from macrosynergy.panel.return_beta import return_beta
benchmark_return = "USD_FXXR_NSA"
data_hedge = return_beta(df=data, xcat='FXXR_NSA', cids=cids,
benchmark_return=benchmark_return, start='2010-01-01',
end='2020-10-30',
blacklist=black, meth='ols', oos=True,
refreq='w', min_obs=24, hedged_returns=True)
print(df_hedge)
beta_display(df_hedge=df_hedge, subplots=False)view_timelines(data, xcats=['FXXR_NSA','FXCRY_NSA'], cids=cids[0],
size=(10, 5), title='AUD Return and Carry')The SignalReturnRelations class analyses and visualises signal and return series.
from macrosynergy.signal.signal_return import SignalReturnRelations
srn = SignalReturnRelations(data, ret="EQXR_NSA", sig="EQCRY_NSA", rival_sigs=None,
sig_neg=True, cosp=True, freq="M", start="2002-01-01")
srn.summary_table()In the creation of the class you can also indicate rival signals for basic relational statistics.
r_sigs = [ "EQCRR_NSA"]
srn = SignalReturnRelations(data, "EQXR_NSA", sig="EQCRY_NSA", rival_sigs=r_sigs,
sig_neg=True, cosp=True, freq="M", start="2002-01-01")
df_sigs = srn.signals_table(sigs=['EQCRY_NSA_NEG', 'EQCRR_NSA_NEG'])
df_sigs_all = srn.signals_table()Using the class you can plot accuracy bars between returns and signals.
srn.accuracy_bars(type="signals", title="Accuracy measure between target return, EQXR_NSA,"
" and the respective signals, ['EQCRY_NSA_NEG', "
" 'EQCRR_NSA_NEG'].")The NaivePnL class computes Pnls with limited signal options and disregarding transaction costs.
from macrosynergy.pnl.naive_pnl import NaivePnL
pnl = NaivePnL(data, ret="EQXR_NSA", sigs=["CRY", "GROWTH"], cids=cids,
start="2000-01-01", bms=["EUR_EQXR_NSA", "USD_EQXR_NSA"])You can then make the pnl and see a list of key pnl statistics.
pnl.make_pnl(
sig="GROWTH", sig_op="zn_score_pan", sig_neg=True, rebal_freq="monthly",
vol_scale=5, rebal_slip=1, min_obs=250, thresh=2)
df_eval = pnl.evaluate_pnls(
pnl_cats=["PNL_GROWTH_NEG"], start="2015-01-01", end="2020-12-31")The official documentation can be found at our documentation website: docs.macrosynergy.com.
We use "code-as-documentation" to ensure that our documentation is always up-to-date. If you find any issues with the documentation, please raise an issue on our GitHub repository.
For the most common issues, such as incorrect credentials, invalid certificates etc, the package will raise an exception with a helpful error message.
If you find that the package raises an HTTPConnection/HTTPSConnectionPool error, please check your proxy settings. In scenarios where an error is raised while running check_connection() (or another download), the error is raised with context to the OAuth token request (to "https://authe.jpmchase.com/as/token.oauth2").
You would most likely need to pass your proxy settings to the JPMaQSDownload object, as shown in the Connecting via a proxy server section.
If you are accessing DataQuery from an institutional/enterprise network, please contact your IT department to ensure that you have the correct proxy settings.
For organizations using ZScaler - you may have to manually add the ZScaler certificates (copy-pasta) to the certifi certificate store (typically called cacert.pem). You can find the location of the certifi certificate store by running the following in your Python environment:
import certifi
print(certifi.where())Please check the documentation for the function on our documentation website,
and ensure you are using the latest version of the package.
If you are still having issues, please raise an issue on our GitHub repository.
Please include a minimal reproducible example, and the output of pip freeze in your issue.
Please raise an issue, and title it "Feature Request: [your feature request]".
Currently, we do not allow a pull request to be created by users outside of the Macrosynergy team. If you'd like to contribute, please create a fork of the repository, and create a pull request from your fork.