IEX Implementation

[rgkimball]

IEX recently announced the first version of a free market data API used to access a wide variety of data useful for traders, consisting of:

• Highest volume stocks in the most recent trading session (Tops API; allows for subset selection)
• Live prices (Last API, also available in websockets)
• Trade volume by venue for the most recent trading session (Market API; separately reported by tape)
• Market Records, by trade volume, symbol volume, route volume and notional value (Stats API)
• Monthly and Daily historical summaries of trade volume (Stats API; separately reported by venue, symbol market cap, and order size)
• Eligible symbols on IEX's platform (via Ref Data API)

This pull request implements all of the services available in their initial launch, except for the live prices via socket, and makes them available in DataFrames via pandas_datareader. Note that the API will return None for weekends and market holidays, and reports a special field for half days. Some examples:

import pandas_datareader as pdr

# Retrieve highest volume stocks for the previous session
tops = pdr.DataReader(source="iex-tops")

# Get most recent price (live)
px = pdr.get_last_iex("AAPL")
px = pdr.get_last_iex(["GOOG", "IBM"])

# Retrieve list of stocks that trade on IEX:
symbols = pdr.get_iex_symbols()

# Retrieve daily historical market volume info
ts = pdr.get_data_iex(start=datetime(2017, 1, 1), end=datetime(2017, 3, 31))

In the coming months, IEX also plans to implement:

• Depth of book (i.e. live level II quotes)
• Company info
• Company financial data
• Historical prices

[jreback]

tests would be nice

[rgkimball]

Added some test cases

[rgkimball]

Rebased for some of gfyoung's recent fixes and cleaned up the formatting.

[rgkimball]

Rebased on 0.4.0 and implemented the newly launched DEEP API. Support implemented for:

• Retrieving live depth of book data (i.e. Level II quotes) for any stock (deep and deep/book)
• Retrieving recent executions with trade size/price, and flags for odd lots, after-market trades, rule 611 exemptions (via deep/trades)
• Check whether the market is open (i.e. deep/system-event)
• Checking trading status of a stock or array of stocks (local and market-wide failures, circuit breakers and IPO status; via deep/trading-status, deep/op-halt-status and ssr-status)
• Retrieving a list of trade breaks (via deep/trade-breaks)

All associated documentation is linked above. Noting that Travis failed due to Yahoo's iCharts API currently being down. Can a maintainer please restart the build when we have confirmation that it is back up? Thank you!

[jreback]

your tests are failing because of importing:

from datetime import (datetime, timezone)

timezone is not a module. you can use pytz directly, or simply use pandas timezone manip (which is vectorized).

[rgkimball]

Ah, thank you. Non-Yahoo tests are passing now.

[jreback]

I would prob do these inside each function this is odd to import everything (though that's the existing style)

e.g.

def get_tops_iex(....):
    from pandas_datareader.iex.tops import TopsReader
    return TopsReader(....)

further would be nice to have a docstring here

[rgkimball]

Agreed, no need to pollute the namespace further even for consistency's sake. Pushed an update with this change and docstrings, but I left imports in for the functions implemented by DataReader.

[jreback]

can you rebase

[davidastephens]

@rgkimball Looks great - can you please rebase and add some docs, then we can merge.

[bashtage]

Can you rebase on master so we can see if things pass.