Python for Trading Course - Part 1: Market Data with OpenBB

What is OpenBB

OpenBB is an open-source platform that integrates more than 100 different data sources, across all asset classes, from stocks, options, crypto assets, forex, macroeconomic data, and much more. It offers access options from a Bloomberg-style app, an old Bloomberg-style CLI console, and leaves the door open to Python programmers through its openbb package.

Historically, accessing financial data was a considerable challenge—it was expensive, inaccessible, and required too much time. With OpenBB, they simplify this entire process by centralizing all data sources in a single point. Using a single library to obtain all our data in real time.

Installing the necessary libraries

Using the Dockerized stack we created

GitHub - quantarmyz/quantstack: QUANTARMY quantstack | One-click quantitative envoirment based in docker, debian, conda, jupyterlabs,zipline-reloaded and arcticdb

QUANTARMY quantstack | One-click quantitative envoirment based in docker, debian, conda, jupyterlabs,zipline-reloaded and arcticdb - quantarmyz/quantstack

GitHubquantarmyz

we load the environment we use from a terminal, and run the following command, which will install the Python gateway that interconnects us with OpenBB, allowing us to work with all the power of OpenBB from Python in an extremely simple way.

!pip install openbb[all]

First Use of OpenBB

Once the library is downloaded, loading the library in Python for the first time will download and install some additional packages in the background.

from openbb import obb
obb.user.preferences.output_type = 'dataframe'

Once finished, it is recommended to restart the kernel, and the library will be fully functional.

Getting Stock Data

Comparing Fundamental Data

obb.equity.fundamental.metrics("AAPL,MSFT").T

Price Data

data = obb.equity.price.historical("SPY")

If we want to change the timeframe, we do it through the interval parameter:

• 1m = One Minute
• 1h = One Hour
• 1d = One Day
• 1W = One Week
• 1M = One Month

In this case, we leave it by default, on a daily timeframe.

5227 rows × 5 columns

Comparing Sectors Fundamentally

obb.equity.compare.groups(group="industry",metric="valuation").T

14 rows × 145 columns

Derivatives: Data from a Futures Curve

data = obb.derivatives.futures.curve(symbol="VX")

Plotting the curve

import pandas as pd
data.index = pd.to_datetime(data.expiration)
data.plot()

Download and plot different contracts

expirations = [
    "2024-12",
    "2025-12",
    "2026-12",
    "2027-12",
    "2028-12",
    "2029-12",
    "2030-12",
]

contracts = []
for expiration in expirations:
    df = (
        obb
        .derivatives
        .futures
        .historical(
            symbol="CL",
            expiration=expiration,
            start_date="2020-01-01",
            end_date="2022-12-31"
        )
    ).rename(columns={
        "close": expiration
            })
    contracts.append(df[expiration])

historical = (
    pd
    .DataFrame(contracts)
    .transpose()
    .dropna()
)

historical.plot()

Getting an Options Chain

Regarding the options chain, we will cover it separately in a future article, exploring all the possibilities that OpenBB provides. For now, to snapshot the current chain of a specific ticker, it would be done as follows:

chains = obb.derivatives.options.chains(symbol="SPY")

9868 rows × 29 columns

Downloading a specific option

If within the chain we want to download a specific contract, we must treat it as if we were downloading a stock ticker, but the symbol is the option identifier. As follows:

data = obb.equity.price.historical(
    symbol="SPY241220C00550000",
        provider="yfinance")

246 rows × 7 columns

Forex

obb.currency.price.historical("usdjpy", provider="yfinance").to_df()

Data Validation

As an extra, we are going to compare data from different providers to get an approximation of the quality of the data we are managing. Data is the input of the process, and low-quality raw material will generate low-quality models. For this test, we will verify the volume on the QQQ ticker (Nasdaq Composite ETF).

To validate that the data from the different providers is correct, we should obtain the same volume for all providers.

First, let's collect the necessary data.

yahoo = obb.equity.price.historical("QQQ", provider="yfinance").to_df()
alphavantage = obb.equity.price.historical("QQQ", provider="alpha_vantage").to_df()
intrinio = obb.equity.price.historical("QQQ", provider="intrinio").to_df()
fmp = obb.equity.price.historical("QQQ", provider="fmp").to_df()
polygon = obb.equity.price.historical("QQQ", provider="polygon").to_df()

Using the following code, we download the QQQ ticker from different providers in Pandas DataFrame format.

Now we process the data, creating a new dataframe with the last 10 cases of the volume from the different providers and removing NA values.

compare = pd.DataFrame()
compare["AV Volume"] = alphavantage["volume"].tail(10)
compare["FMP Volume"] = fmp["volume"].tail(10)
compare["Intrinio Volume"] = intrinio["volume"].tail(10)
compare["Yahoo Volume"] = yahoo["volume"].tail(10)
compare["Polygon Volume"] = polygon["volume"].tail(10)

compare.dropna(how="any")

As a conclusion, we can see that they do not always match, so everyone must run their own procedures to determine which price is real and which is not.

Conclusions

These short articles aim to get straight to the point in a concentrated way. Through this article, you can access and download time series of different asset classes, to later manage them in our datalake, or in whatever independent way you need.