Datasets
Alternative Data
Introduction
This page explains how to request, manipulate, and visualize historical alternative data. This tutorial uses the VIX Daily Price dataset from the CBOE as the example dataset.
Create Subscriptions
Follow these steps to subscribe to an alternative dataset from the Dataset Market:
- Load the required assembly files and data types.
- Load the dynamic link library (DLL) of the dataset.
- Create a
QuantBook
. - Call the
AddData
add_data
method with the dataset class, a ticker, and a resolution and then save a reference to the alternative dataSymbol
.
#load "../Initialize.csx" #load "../QuantConnect.csx" #r "../Microsoft.Data.Analysis.dll" using QuantConnect; using QuantConnect.Data; using QuantConnect.Algorithm; using QuantConnect.Research; using QuantConnect.DataSource; using Microsoft.Data.Analysis;
To load the DLL of any dataset, type:
#r "../QuantConnect.DataSource.<nameOfAlternativeDatasetClass>.dll"
For example, to load the DLL of the CBOE dataset, type:
#r "../QuantConnect.DataSource.CBOE.dll"
var qb = new QuantBook();
qb = QuantBook()
var vix = qb.AddData<CBOE>("VIX", Resolution.Daily).Symbol; var v3m = qb.AddData<CBOE>("VIX3M", Resolution.Daily).Symbol;
vix = qb.add_data(CBOE, "VIX", Resolution.DAILY).symbol v3m = qb.add_data(CBOE, "VIX3M", Resolution.DAILY).symbol
To view the arguments that the AddData
add_data
method accepts for each dataset, see the dataset listing.
If you don't pass a resolution argument, the default resolution of the dataset is used by default. To view the supported resolutions and the default resolution of each dataset, see the dataset listing.
Get Historical Data
You need a subscription before you can request historical data for a dataset. On the time dimension, you can request an amount of historical data based on a trailing number of bars, a trailing period of time, or a defined period of time. On the dataset dimension, you can request historical data for a single dataset subscription, a subset of the dataset subscriptions you created in your notebook, or all of the dataset subscriptions in your notebook.
Trailing Number of Bars
To get historical data for a number of trailing bars, call the History
history
method with the Symbol
object(s) and an integer.
// Slice objects var singleHistorySlice = qb.History(vix, 10); var subsetHistorySlice = qb.History(new[] {vix, v3m}, 10); var allHistorySlice = qb.History(10); // CBOE objects var singleHistoryDataObjects = qb.History<CBOE>(vix, 10); var subsetHistoryDataObjects = qb.History<CBOE>(new[] {vix, v3m}, 10); var allHistoryDataObjects = qb.History<CBOE>(qb.Securities.Keys, 10);
# DataFrame single_history_df = qb.History(vix, 10) subset_history_df = qb.History([vix, v3m], 10) all_history_df = qb.History(qb.Securities.Keys, 10) # Slice objects all_history_slice = qb.History(10) # CBOE objects single_history_data_objects = qb.History[CBOE](vix, 10) subset_history_data_objects = qb.History[CBOE]([vix, v3m], 10)
all_history_data_objects = qb.History[CBOE](qb.Securities.Keys, 10)
The preceding calls return the most recent bars, excluding periods of time when the exchange was closed.
Trailing Period of Time
To get historical data for a trailing period of time, call the History
history
method with the Symbol
object(s) and a TimeSpan
timedelta
.
// Slice objects var singleHistorySlice = qb.History(vix, TimeSpan.FromDays(3)); var subsetHistorySlice = qb.History(new[] {vix, v3m}, TimeSpan.FromDays(3)); var allHistorySlice = qb.History(10); // CBOE objects var singleHistoryDataObjects = qb.History<CBOE>(vix, TimeSpan.FromDays(3));
var subsetHistoryDataObjects = qb.History<CBOE>(new[] {vix, v3m}, TimeSpan.FromDays(3));
var allHistoryDataObjects = qb.History<CBOE>(TimeSpan.FromDays(3));
# DataFrame single_history_df = qb.History(vix, timedelta(days=3)) subset_history_df = qb.History([vix, v3m], timedelta(days=3)) all_history_df = qb.History(qb.Securities.Keys, timedelta(days=3)) # Slice objects all_history_slice = qb.History(timedelta(days=3)) # CBOE objects single_history_data_objects = qb.History[CBOE](vix, timedelta(days=3))
subset_history_data_objects = qb.History[CBOE]([vix, v3m], timedelta(days=3))
all_history_data_objects = qb.History[CBOE](qb.Securities.Keys, timedelta(days=3))
The preceding calls return the most recent bars or ticks, excluding periods of time when the exchange was closed.
Defined Period of Time
To get historical data for a specific period of time, call the History
history
method with the Symbol
object(s), a start DateTime
datetime
, and an end DateTime
datetime
. The start and end times you provide are based in the notebook time zone.
var startTime = new DateTime(2021, 1, 1); var endTime = new DateTime(2021, 3, 1); // Slice objects var singleHistorySlice = qb.History(vix, startTime, endTime); var subsetHistorySlice = qb.History(new[] {vix, v3m}, startTime, endTime); var allHistorySlice = qb.History(startTime, endTime); // CBOE objects var singleHistoryDataObjects = qb.History<CBOE>(vix, startTime, endTime);
var subsetHistoryDataObjects = qb.History<CBOE>(new[] {vix, v3m}, startTime, endTime);
var allHistoryDataObjects = qb.History<CBOE>(qb.Securities.Keys, startTime, endTime);
start_time = datetime(2021, 1, 1) end_time = datetime(2021, 3, 1) # DataFrame single_history_df = qb.History(vix, start_time, end_time) subset_history_df = qb.History([vix, v3m], start_time, end_time) all_history_df = qb.History(qb.Securities.Keys, start_time, end_time) # Slice objects all_history_slice = qb.History(start_time, end_time) # CBOE objects single_history_data_objects = qb.History[CBOE](vix, start_time, end_time)
subset_history_data_objects = qb.History[CBOE]([vix, v3m], start_time, end_time)
all_history_data_objects = qb.History[CBOE](qb.Securities.Keys, start_time, end_time)
The preceding calls return the bars or ticks that have a timestamp within the defined period of time.
If you do not pass a resolution to the History
history
method, the History
history
method uses the resolution that the AddData
add_data
method used when you created the subscription.
Wrangle Data
You need some historical data to perform wrangling operations. The process to manipulate the historical data depends on its data type. To display pandas
objects, run a cell in a notebook with the pandas
object as the last line. To display other data formats, call the print
method.
You need some historical data to perform wrangling operations. Use LINQ to wrangle the data and then call the Console.WriteLine
method in a Jupyter Notebook to display the data. The process to manipulate the historical data depends on its data type.
DataFrame Objects
If the History
history
method returns a DataFrame
, the first level of the DataFrame
index is the encoded dataset Symbol and the second level is the EndTime
end_time
of the data sample. The columns of the DataFrame
are the data properties.
To select the historical data of a single dataset, index the loc
property of the DataFrame
with the dataset Symbol
.
all_history_df.loc[vix] # or all_history_df.loc['VIX']
To select a column of the DataFrame
, index it with the column name.
all_history_df.loc[vix]['close']
If you request historical data for multiple tickers, you can transform the DataFrame
so that it's a time series of close values for all of the tickers. To transform the DataFrame
, select the column you want to display for each ticker and then call the unstack method.
all_history_df['close'].unstack(level=0)
The DataFrame
is transformed so that the column indices are the Symbol
of each ticker and each row contains the close value.
The historical data methods don't return DataFrame objects, but you can create one for efficient vectorized data wrangling.
using Microsoft.Data.Analysis; var columns = new DataFrameColumn[] { new PrimitiveDataFrameColumn("Time", history.Select(x => x[vix].EndTime)), new DecimalDataFrameColumn("VIX Open", history.Select(x => x[vix].Open)), new DecimalDataFrameColumn("VIX High", history.Select(x => x[vix].High)), new DecimalDataFrameColumn("VIX Low", history.Select(x => x[vix].Low)), new DecimalDataFrameColumn("VIX Close", history.Select(x => x[vix].Close)) }; var df = new DataFrame(columns); df
To select a particular column of the DataFrame, index it with the column name.
df["VIX close"]
Slice Objects
If the History
history
method returns Slice
objects, iterate through the Slice
objects to get each one. The Slice
objects may not have data for all of your dataset subscriptions. To avoid issues, check if the Slice
contains data for your ticker before you index it with the dataset Symbol
.
foreach (var slice in allHistorySlice) { if (slice.ContainsKey(vix)) { var data = slice[vix]; } }
for slice in all_history_slice: if slice.contains_key(vix): data = slice[vix]
Plot Data
You need some historical alternative data to produce plots. You can use many of the supported plotting librariesPlot.NET package to visualize data in various formats. For example, you can plot candlestick and line charts.
Candlestick Chart
You can only create candlestick charts for alternative datasets that have open, high, low, and close properties.
Follow these steps to plot candlestick charts:
- Get some historical data.
- Import the
plotly
Plot.NET
library. - Create a
Candlestick
. - Create a
Layout
. - Create a
Figure
. - Assign the
Layout
to the chart. - Show the
Figure
.
history = qb.history(vix, datetime(2021, 1, 1), datetime(2021, 2, 1)).loc[vix]
var history = qb.History<CBOE>(vix, new DateTime(2021, 12, 1), new DateTime(2021, 12, 31));
import plotly.graph_objects as go
#r "../Plotly.NET.dll" using Plotly.NET; using Plotly.NET.LayoutObjects;
candlestick = go.Candlestick(x=history.index, open=history['open'], high=history['high'], low=history['low'], close=history['close'])
var chart = Chart2D.Chart.Candlestick<decimal, decimal, decimal, decimal, DateTime, string>( history.Select(x => x.Open), history.Select(x => x.High), history.Select(x => x.Low), history.Select(x => x.Close), history.Select(x => x.EndTime) );
layout = go.Layout(title=go.layout.Title(text='VIX from CBOE OHLC'), xaxis_title='Date', yaxis_title='Price', xaxis_rangeslider_visible=False)
LinearAxis xAxis = new LinearAxis(); xAxis.SetValue("title", "Time"); LinearAxis yAxis = new LinearAxis(); yAxis.SetValue("title", "Price ($)"); Title title = Title.init($"{vix} OHLC"); Layout layout = new Layout(); layout.SetValue("xaxis", xAxis); layout.SetValue("yaxis", yAxis); layout.SetValue("title", title);
fig = go.Figure(data=[candlestick], layout=layout)
chart.WithLayout(layout);
fig.show()
HTML(GenericChart.toChartHTML(chart))
Candlestick charts display the open, high, low, and close prices of the alternative data.
Line Chart
Follow these steps to plot line charts using built-in methodsPlotly.NET
package:
- Get some historical data.
- Select the data to plot.
- Call the
plot
method on thepandas
object. - Create
Line
charts. - Create a
Layout
. - Combine the charts and assign the
Layout
to the chart. - Show the plot.
history = qb.history([vix, v3m], datetime(2021, 1, 1), datetime(2021, 2, 1))
var history = qb.History<CBOE>(new [] {vix, v3x}, new DateTime(2021, 1, 1), new DateTime(2021, 2, 1));
values = history['close'].unstack(0)
values.plot(title = 'Close', figsize=(15, 10))
var chart1 = Chart2D.Chart.Line<DateTime, decimal, string>( history.Select(x => x[vix].EndTime), history.Select(x => x[vix].Close), Name: $"{vix}" ); var chart2 = Chart2D.Chart.Line<DateTime, decimal, string>( history.Select(x => x[v3m].EndTime), history.Select(x => x[v3m].Close), Name: $"{v3m}" );
LinearAxis xAxis = new LinearAxis(); xAxis.SetValue("title", "Time"); LinearAxis yAxis = new LinearAxis(); yAxis.SetValue("title", "Price ($)"); Title title = Title.init($"{vix} & {v3m} Close Price"); Layout layout = new Layout(); layout.SetValue("xaxis", xAxis); layout.SetValue("yaxis", yAxis); layout.SetValue("title", title);
var chart = Plotly.NET.Chart.Combine(new []{chart1, chart2}); chart.WithLayout(layout);
plt.show()
HTML(GenericChart.toChartHTML(chart))
Line charts display the value of the property you selected in a time series.