Eikon API Proxy
Thomson Reuters
Eikon Data API
The Thomson Reuters Eikon Data API Proxy is a new modern desktop API geared towards Quants and Data Scientists.
The API Proxy is completely integrated in Eikon 4 (Windows) or can be installed as a standalone package in Mac.
Complementing the users' workflow.
The Data Scientist's Python Ecosystem¶
Imports¶
In [1]:
import eikon as ek # Data Source
import pandas as pd # Data Management
import cufflinks as cf # Data Visualization
A Secure Approach¶
The app id can be obtained from the API Key Generator
In [4]:
ek.set_app_id('AD2833XXXXXXXX')
Symbology
Conversion
We can convert across all types of symbology
In [21]:
sedols = ['B1YW440', '0673123', 'B02J639', 'B1XZS82', '0045614', '0053673',
'0989529', '0216238', '0969703', '0263494']
In [31]:
ek.get_symbology(sedols,to_symbol_type=['ISIN','RIC'])
Out[31]:
News
Getting the top news for Apple where Trump is mentioned
In [33]:
ek.get_news_headlines('R:.SPX AND "Comey" AND Language:LEN',count=5)
Out[33]:
And then we can retrieve the actual story. The stories are delivered in HTML - which means we can display them formatted in our Notebook.
In [34]:
from IPython.display import display,HTML
In [36]:
display(HTML(ek.get_news_story('urn:newsml:reuters.com:20170510:nL4N1IC4KW')))
Analyzing 2 years
Of Microsoft Data
Using ek.get_timeseries to retrieve 2 years of Microsoft Data
In [5]:
msft=ek.get_timeseries('MSFT.O',fields='CLOSE',start_date='2015-05-04',end_date='2017-05-04')
In [6]:
msft.head()
Out[6]:
Getting a best fit line
In [8]:
msft.iplot(kind='lines',color='blue',width=2,bestfit=True,title='Bestfit')
Adding a Quarterly Results Event to a Microsoft Chart
In [9]:
# All annotations have basic support for HTML tags
ann={'2016-01-29':'Microsoft Reports Quarterly Results Above Expectations' \
'<BR>Lowering Outlook, Price Target and Rating.'}
In [10]:
msft.iplot(kind='lines',annotations=ann,textangle=0,title='Microsoft - Estimate Revisions')
How sensitive is the price of Microsoft to Dividend Payments?¶
We would like to retrieve Net (Adjusted) Dividends and overlay this information over the price chart of Microsoft.
In [11]:
# We will retreive Net and Gross Dividend for From May 2015 to April 2017
fields=['TR.DivExDate','TR.DivAdjustedGross','TR.DivAdjustedNet']
divs,err=ek.get_data('MSFT.O',fields=fields,
parameters={'SDate':'05May15','EDate':'30Apr17','DateType':'AD'})
Plotting the results
In [14]:
cf.tools.merge_figures([msft_price,msft_divs]).iplot(legend=False)
Performance
Comparison
Has Microsoft outperformed the SPX index?¶
We will retrieve data from 2015 for both Microsoft and SPX
In [14]:
performance=ek.get_timeseries(['MSFT.O','.SPX'],
fields='CLOSE',
start_date='2015-05-04',
end_date='2017-05-04')
In [15]:
performance.head()
Out[15]:
Normalizing (Percent change from day 0) the values for both instruments and charting the performance.
In [16]:
performance.normalize().iplot(kind='spread',title='Microsoft Performance vs SPX',
width=2,showgrid=False)
Returns
Heatmap
We want to retrieve the 1 month returns for all the companies in STOXX and display a heathmap of the mean returns by country and sector.¶
We will get the 600 companies and the 1 month return for each of them
In [15]:
# Support for chains
df,err=ek.get_data('0#.STOXX',['TR.HQCountryCode','TR.TRBCEconomicSector','TR.TotalReturn1Mo'])
In [16]:
df.head()
Out[16]:
Group them by Country/Sector
In [17]:
group=df.groupby(['Country ISO Code of Headquarters','TRBC Economic Sector Name'])
matrix=group.mean().unstack()['1 Month Total Return']
And plot the results
In [18]:
matrix.iplot(kind='heatmap',colorscale='rdylgn',margin={'l':200},
title='1 Month Mean Returns in Europe')
Growth
Estimate Changes
We want to retrieve the Percent Change (Growth) of Estimates across multiple Fiscal Years¶
Getting the 4 Year YoY Estimate change on EPS and Revenue for Amazon ¶
In [30]:
amz_chg=get_estimates('AMZN.O',fields=['EPS','Revenue'],periods=4)
In [31]:
amz_chg
Out[31]:
In [32]:
amz_chg.iplot(kind='bar',title='Amazon Estimates YoY Changes',colorscale='-brbg',
yTitle='Percent Change',theme='ggplot')
How good or bad are these Estimate Changes compared to Amazon's Peer Group ?¶
We first retrieve the peers of Amazon (including Amazon)
In [33]:
peers,err=ek.get_data(['PEERS(AMZN.O)','AMZN.O'],['TR.CommonName','TR.CompanyMarketCap'])
And get the top 10 by Market Cap
In [34]:
peers=peers.sort_values(by='Company Market Cap',ascending=False)[:10]
peers
Out[34]:
In [38]:
peers_df.iplot(kind='bar',colorscale='-purples',title='Amazon Peers - EPS Est % Change YoY',
yTitle='Percentage')
Growth
Ideas
Screening for TECH companies in Russell 1000 Growth Index with a Market Cap between $5B and $20B and an Analyst Revision greater than 80/100.¶
First we define the Universe over which the screening will be performed
In [40]:
# From the Russell 1000 Growth Index as strarting universe
# With a Market Cap between 5B and 20B USD
# With Analyst recommendations greater than 80 and Price to Mid Term component greater than 30
# And EPS >=10 and Revenue >=5 Smart Estimates (All in USD)
universe='SCREEN(U(IN(indices(4387903/*Russell 1000 Growth*/))), ' \
'TR.CompanyMarketCap>=5000000000, TR.CompanyMarketCap<20000000000, ' \
'TR.ARM100Region>=80, TR.PriceMoMidTermComponent>=30, ' \
'TR.EPSSmartEstLastYrGrowth>=10, TR.RevenueSmartEstLastYrGrowth>=5, CURN=USD)'
fields=['TR.CommonName','TR.PricePctChg52W',
'TR.CompanyMarketCap','TR.Revenue(Period:FY0)',
'TR.ARM100Region','TR.TRBCEconomicSector']
In [41]:
df,err=ek.get_data(universe,fields=fields)
We get the top 50 companies by Market Cap
In [42]:
df=df.sort_values(by='Company Market Cap',ascending=False)[:50]
df.head()
Out[42]:
In [44]:
df.iplot(kind='bubble',x='52-week Price PCT Change',y='Revenue',size='Company Market Cap',
text='text',categories='TRBC Economic Sector Name',xTitle='52-week Price PCT Change',
yTitle='Revenue',zerolinewidth=2,zerolinecolor='darkgrey',colorscale='set3')
We can add another axis (z) and see them in a 3D form
In [45]:
df.iplot(kind='bubble3d',x='52-week Price PCT Change',y='Revenue',z='Analyst Revisions Model Region Rank',
size='Company Market Cap',text='Company Common Name',categories='TRBC Economic Sector Name',
xTitle='52-week Price PCT Change',yTitle='Revenue',zTitle='Analyst Recs',zerolinewidth=3,
zerolinecolor='darkgrey',colorscale='set3',width=0)
Analyst
Recommendations
What do Analysts think
of the FTSE these days?
We retrieve the analysts' recommendations for the FTSE
In [56]:
recs,err=ek.get_data('0#.FTSE',['TR.CompanyName','TR.NumOfBuy','TR.NumOfHold','TR.NumOfSell'])
In [58]:
recs.head()
Out[58]:
And we chart the results for all companies
In [60]:
recs.iplot(kind='barh',barmode='stack',margin={'l':300},dimensions=(1000,600),
colors=['green','blue','pink'])
Easier to digest by sectors?¶
Plotting the sector results
In [64]:
recs.iplot(kind='barh',barmode='stack',margin={'l':200},colors=['green','blue','pink'],
title='FTSE Analyst Recommendations by Sector')
Returns
Distribution
US Presidents
The first 100 days
We will analyze the first 100 days of each of the last 5 Presidents of the United States and look at the distribution of returns for the SPX¶
Presidential Inauguration Dates
In [65]:
potus={'Barack Obama': '2009-01-20',
'Donald J. Trump': '2017-01-20',
'George Bush': '1989-01-20',
'William J. Clinton': '1993-01-20',
'George W. Bush': '2001-01-20'}
In [69]:
potus_df={}
for p in potus:
# Getting the SPX closing values for the first 100 days
df=ek.get_timeseries('.SPX','CLOSE',
start_date=potus[p],
end_date=cf.date_tools.getDatefromDate(potus[p],100,'%Y-%m-%d'))
potus_df[p]=df['CLOSE'].values
In [74]:
df.pct_change(1).iplot(kind='histogram',colorscale='set1',width=0.5,legend='bottom',
title='Potus First 100 Days<br>1d Returns Distribution')
In [76]:
df.pct_change(1).iplot(kind='histogram',colorscale='set1',width=0.5,legend='bottom',
title='Potus First 100 Days<br>1d Returns Distribution',subplots=True,
shared_xaxes=True)
In [77]:
df.normalize().iplot(colorscale='set1',width=2,legend='bottom',title='First 100 Days POTUS Returns (SPX)')
Timeseries
Analyzing USD/GBP
In [153]:
# 7 Years of OHLC Weekly data
gbp=ek.get_timeseries('gbp=',fields='*',start_date='2010-01-01',interval='weekly')
Optimized for immediate charting
In [109]:
qf.iplot()
SPX Long Term
In [110]:
spx=ek.get_timeseries('.spx',fields='*',start_date='2010-01-01',interval='weekly')
In [119]:
qf.iplot()
Technical Analysis over Apple
In [121]:
aapl=ek.get_timeseries('aapl.o',fields='*',start_date='2015-01-01')
In [141]:
qf=cf.QuantFig(aapl,legend='top',name='Apple')
In [143]:
qf.add_volume(colorchange=True)
qf.add_sma([21,55,144],colors=['green','orange','blue'],width=1)
Simple Moving Averages
In [144]:
qf.iplot()
Adding Bollinger Bands, Volume, MACD and RSI
In [152]:
qf.iplot(showlegend=False,title='Apple')
In [ ]:
Vessel
Search
Searching for Vessels (Dry Bulk) transporting Iron Ore¶
Retrieving all Dry Bulk (Bulkers) currently transporting Iron Ore alongside with its geographical coordinates and tonnage information.
In [12]:
fields=['TR.AssetName','TR.AssetLocationLatitude','TR.AssetLocationLongitude',
'TR.AssetDestinationPort','TR.AssetDWT']
In [6]:
df,err=ek.get_data('SCREEN(U(IN(lists(337598580))))',fields)
In [7]:
df.head()
Out[7]:
Plotting all vessels Colorscale indicated Dead Weight Tonnage
In [11]:
fig.iplot()