Pandas TA (pandas_ta) Studies for Custom Technical Analysis

Topics

What is a Pandas TA Study?
- Builtin Studies: AllStudy and CommonStudy
- Creating Studies
Watchlist Class
- Study Management and Execution
- NOTE: The watchlist module is independent of Pandas TA. To easily use it, copy it from your local pandas_ta installation directory into your project directory.
Indicator Composition/Chaining for more Complex Studies
- Comprehensive Example: MACD and RSI Momo with BBANDS and SMAs 50 & 200 and Cumulative Log Returns

In [1]:

%matplotlib inline
import datetime as dt

from tqdm import tqdm

import pandas as pd
import pandas_ta as ta
from alphaVantageAPI.alphavantage import AlphaVantage  # pip install alphaVantage-api

from watchlist import Watchlist # Is this failing? If so, copy it locally. See above.

print(f"\nPandas TA v{ta.version}\nTo install the Latest Version:\n$ pip install -U git+https://github.com/twopirllc/pandas-ta\n")
%pylab inline

Pandas TA v0.3.32b0
To install the Latest Version:
$ pip install -U git+https://github.com/twopirllc/pandas-ta

Populating the interactive namespace from numpy and matplotlib

What is a Pandas TA Study?

A Study is a simple way to name and group TA indicators. Technically, a Study is a simple Data Class to contain list of indicators and their parameters. Note: Study is experimental and subject to change. Pandas TA comes with two basic Studies: AllStudy and CommonStudy.

Study Requirements:

name: Some short memorable string. Note: Case-insensitive "All" is reserved.
ta: A list of dicts containing keyword arguments to identify the indicator and the indicator's arguments

Optional Requirements:

description: A more detailed description of what the Study tries to capture. Default: None
created: At datetime string of when it was created. Default: Automatically generated.

Things to note:

A Study will fail when consumed by Pandas TA if there is no {"kind": "indicator name"} attribute.

Builtin Examples

All

Default Values

In [2]:

AllStudy = ta.AllStudy
print("name =", AllStudy.name)
print("description =", AllStudy.description)
print("created =", AllStudy.created)
print("ta =", AllStudy.ta)

name = All
description = All the indicators with their default settings. Pandas TA default.
created = Sunday January 23, 2022, NYSE: 11:36:06, Local: 15:36:06 PST, Day 23/365 (6.00%)
ta = None

Common

Default Values

In [3]:

CommonStudy = ta.CommonStudy
print("name =", CommonStudy.name)
print("description =", CommonStudy.description)
print("created =", CommonStudy.created)
print("ta =", CommonStudy.ta)

name = Common Price and Volume SMAs
description = Common Price SMAs: 10, 20, 50, 200 and Volume SMA: 20.
created = Sunday January 23, 2022, NYSE: 11:36:06, Local: 15:36:06 PST, Day 23/365 (6.00%)
ta = [{'kind': 'sma', 'length': 10}, {'kind': 'sma', 'length': 20}, {'kind': 'sma', 'length': 50}, {'kind': 'sma', 'length': 200}, {'kind': 'sma', 'close': 'volume', 'length': 20, 'prefix': 'VOL'}]

In [ ]:

Creating Studies

Studies require a name and an array of dicts containing the "kind" of indicator ("sma") and other potential parameters for ta.

Simple Study A

In [4]:

custom_a = ta.Study(name="A", ta=[{"kind": "sma", "length": 50}, {"kind": "sma", "length": 200}])
custom_a

Out [4]:

Study(name='A', ta=[{'kind': 'sma', 'length': 50}, {'kind': 'sma', 'length': 200}], description='', created='Sunday January 23, 2022, NYSE: 11:36:06, Local: 15:36:06 PST, Day 23/365 (6.00%)')

Simple Study B

In [5]:

custom_b = ta.Study(name="B", ta=[{"kind": "ema", "length": 8}, {"kind": "ema", "length": 21}, {"kind": "log_return", "cumulative": True}, {"kind": "rsi"}, {"kind": "supertrend"}])
custom_b

Out [5]:

Study(name='B', ta=[{'kind': 'ema', 'length': 8}, {'kind': 'ema', 'length': 21}, {'kind': 'log_return', 'cumulative': True}, {'kind': 'rsi'}, {'kind': 'supertrend'}], description='', created='Sunday January 23, 2022, NYSE: 11:36:06, Local: 15:36:06 PST, Day 23/365 (6.00%)')

Bad Study. (Misspelled Indicator)

In [6]:

# Misspelled indicator, will fail later when ran with Pandas TA
custom_run_failure = ta.Study(name="Runtime Failure", ta=[{"kind": "peret_return"}])
custom_run_failure

Out [6]:

Study(name='Runtime Failure', ta=[{'kind': 'peret_return'}], description='', created='Sunday January 23, 2022, NYSE: 11:36:06, Local: 15:36:06 PST, Day 23/365 (6.00%)')

In [ ]:

Study Management and Execution with Watchlist

Initialize AlphaVantage Data Source

In [7]:

AV = AlphaVantage(
    api_key="YOUR API KEY", premium=False,
    output_size='full', clean=True,
    export_path=".", export=True
)
AV

Out [7]:

AlphaVantage(
  end_point:str = https://www.alphavantage.co/query,
  api_key:str = YOUR API KEY,
  export:bool = True,
  export_path:str = .,
  output_size:str = full,
  output:str = csv,
  datatype:str = json,
  clean:bool = True,
  proxy:dict = {}
)

Create Watchlist and set it's 'ds' to AlphaVantage

In [8]:

data_source = "av" # Default
data_source = "yahoo"
watch = Watchlist(["SPY", "IWM"], ds_name=data_source, timed=False)

Info about the Watchlist. Note, the default Study is "All"

In [9]:

watch

Out [9]:

Watch(name='Watch: SPY, IWM', ds_name='yahoo', tickers[2]='SPY, IWM', tf='D', study[5]='Common Price and Volume SMAs')

Help about Watchlist

In [10]:

help(Watchlist)

Help on class Watchlist in module watchlist:

class Watchlist(builtins.object)
 |  Watchlist(tickers: list, tf: str = None, name: str = None, study: pandas_ta.utils._study.Study = None, ds_name: str = 'av', **kwargs)
 |  
 |  # Watchlist Class (** This is subject to change! **)
 |  A simple Class to load/download financial market data and automatically
 |  apply Technical Analysis indicators with a Pandas TA Study.
 |  
 |  Default Study: pandas_ta.CommonStudy
 |  
 |  ## Package Support:
 |  ### Data Source (Default: AlphaVantage)
 |  - AlphaVantage (pip install alphaVantage-api).
 |  - Python Binance (pip install python-binance). # Future Support
 |  - Yahoo Finance (pip install yfinance). # Almost Supported
 |  
 |  # Technical Analysis:
 |  - Pandas TA (pip install pandas_ta)
 |  
 |  ## Required Arguments:
 |  - tickers: A list of strings containing tickers. Example: ["SPY", "AAPL"]
 |  
 |  Methods defined here:
 |  
 |  __init__(self, tickers: list, tf: str = None, name: str = None, study: pandas_ta.utils._study.Study = None, ds_name: str = 'av', **kwargs)
 |      Initialize self.  See help(type(self)) for accurate signature.
 |  
 |  __repr__(self) -> str
 |      Return repr(self).
 |  
 |  indicators(self, *args, **kwargs) -> <built-in function any>
 |      Returns the list of indicators that are available with Pandas Ta.
 |  
 |  load(self, ticker: str = None, tf: str = None, index: str = 'date', drop: list = [], plot: bool = False, **kwargs) -> pandas.core.frame.DataFrame
 |      Loads or Downloads (if a local csv does not exist) the data from the
 |      Data Source. When successful, it returns a Data Frame for the requested
 |      ticker. If no tickers are given, it loads all the tickers.
 |  
 |  ----------------------------------------------------------------------
 |  Data descriptors defined here:
 |  
 |  __dict__
 |      dictionary for instance variables (if defined)
 |  
 |  __weakref__
 |      list of weak references to the object (if defined)
 |  
 |  data
 |      When not None, it contains a dictionary of DataFrames keyed by ticker. data = {"SPY": pd.DataFrame, ...}
 |  
 |  name
 |      The name of the Watchlist. Default: "Watchlist: {Watchlist.tickers}".
 |  
 |  strategy
 |      Sets a valid Strategy. Default: pandas_ta.CommonStrategy
 |  
 |  study
 |      Sets a valid Study. Default: pandas_ta.CommonStudy
 |  
 |  tf
 |      Alias for timeframe. Default: 'D'
 |  
 |  tickers
 |      tickers
 |      
 |      If a string, it it converted to a list. Example: "AAPL" -> ["AAPL"]
 |          * Does not accept, comma seperated strings.
 |      If a list, checks if it is a list of strings.
 |  
 |  verbose
 |      Toggle the verbose property. Default: False

Default Study is "Common"

In [11]:

# No arguments loads all the tickers and applies the Study to each ticker.
# The result can be accessed with Watchlist's 'data' property which returns a 
# dictionary keyed by ticker and DataFrames as values 
watch.load(verbose=True)

[!] Loading All: SPY, IWM
[+] Downloading[yahoo]: SPY[D]
[+] Saving: /Users/kj/av_data/SPY_D.csv
[+] Study: Common Price and Volume SMAs
[i] Indicator arguments: {'timed': False, 'append': True}
[i] Multiprocessing 5 indicators with 7 chunks and 8/8 cpus.

100%|██████████████████████████████████████████████████████████████████████████████████████████████████████████| 5/5 [00:00<00:00, 66.43it/s]

[i] Total indicators: 5
[i] Columns added: 5
[i] Last Run: Sunday January 23, 2022, NYSE: 11:36:11, Local: 15:36:11 PST, Day 23/365 (6.00%)
[+] Downloading[yahoo]: IWM[D]
[+] Saving: /Users/kj/av_data/IWM_D.csv
[+] Study: Common Price and Volume SMAs
[i] Indicator arguments: {'timed': False, 'append': True}
[i] Multiprocessing 5 indicators with 7 chunks and 8/8 cpus.

100%|██████████████████████████████████████████████████████████████████████████████████████████████████████████| 5/5 [00:00<00:00, 62.67it/s]

[i] Total indicators: 5
[i] Columns added: 5
[i] Last Run: Sunday January 23, 2022, NYSE: 11:36:15, Local: 15:36:15 PST, Day 23/365 (6.00%)

In [12]:

", ".join([f"{t}: {d.shape}" for t,d in watch.data.items()])

Out [12]:

'SPY: (7299, 12), IWM: (5449, 12)'

In [13]:

watch.data["SPY"]

Out [13]:

	open	high	low	close	volume	dividends	stock splits	SMA_10	SMA_20	SMA_50	SMA_200	VOL_SMA_20
date
1993-01-29	25.645575	25.645575	25.517985	25.627348	1003200	0.0	0	NaN	NaN	NaN	NaN	NaN
1993-02-01	25.645562	25.809607	25.645562	25.809607	480500	0.0	0	NaN	NaN	NaN	NaN	NaN
1993-02-02	25.791391	25.882527	25.736710	25.864300	201300	0.0	0	NaN	NaN	NaN	NaN	NaN
1993-02-03	25.900745	26.155924	25.882517	26.137697	529400	0.0	0	NaN	NaN	NaN	NaN	NaN
1993-02-04	26.228839	26.301748	25.937205	26.247066	531500	0.0	0	NaN	NaN	NaN	NaN	NaN
...	...	...	...	...	...	...	...	...	...	...	...	...
2022-01-14	461.190002	465.089996	459.899994	464.720001	95849600	0.0	0	469.319998	469.606500	465.989207	438.476059	78776895.0
2022-01-18	459.739990	459.959991	455.309998	456.489990	109709100	0.0	0	467.197998	469.437500	465.813499	438.746869	77486770.0
2022-01-19	458.130005	459.609985	451.459991	451.750000	109357600	0.0	0	464.617999	469.275999	465.510704	438.995167	77597910.0
2022-01-20	453.750000	458.739990	444.500000	446.750000	122379700	0.0	0	462.454999	468.460500	465.099938	439.216139	80226580.0
2022-01-21	445.559998	448.059998	437.950012	437.980011	201913500	0.0	0	459.459000	466.975000	464.544664	439.383706	87377745.0

7299 rows × 12 columns

In [ ]:

In [14]:

watch.load("SPY", plot=True, mas=True)

Out [14]:

[i] Loaded SPY[D]: SPY_D.csv

	open	high	low	close	volume	dividends	stock splits	SMA_10	SMA_20	SMA_50	SMA_200	VOL_SMA_20
date
1993-01-29	25.645575	25.645575	25.517985	25.627348	1003200	0.0	0	NaN	NaN	NaN	NaN	NaN
1993-02-01	25.645562	25.809607	25.645562	25.809607	480500	0.0	0	NaN	NaN	NaN	NaN	NaN
1993-02-02	25.791391	25.882527	25.736710	25.864300	201300	0.0	0	NaN	NaN	NaN	NaN	NaN
1993-02-03	25.900745	26.155924	25.882517	26.137697	529400	0.0	0	NaN	NaN	NaN	NaN	NaN
1993-02-04	26.228839	26.301748	25.937205	26.247066	531500	0.0	0	NaN	NaN	NaN	NaN	NaN
...	...	...	...	...	...	...	...	...	...	...	...	...
2022-01-14	461.190002	465.089996	459.899994	464.720001	95849600	0.0	0	469.319998	469.606500	465.989207	438.476059	78776895.0
2022-01-18	459.739990	459.959991	455.309998	456.489990	109709100	0.0	0	467.197998	469.437500	465.813499	438.746869	77486770.0
2022-01-19	458.130005	459.609985	451.459991	451.750000	109357600	0.0	0	464.617999	469.275999	465.510704	438.995167	77597910.0
2022-01-20	453.750000	458.739990	444.500000	446.750000	122379700	0.0	0	462.454999	468.460500	465.099938	439.216139	80226580.0
2022-01-21	445.559998	448.059998	437.950012	437.980011	201913500	0.0	0	459.459000	466.975000	464.544664	439.383706	87377745.0

7299 rows × 12 columns

In [ ]:

Easy to swap Studies and run them

Running Simple Study A

In [15]:

# Load custom_a into Watchlist and verify
watch.study = custom_a
watch.study

Out [15]:

Study(name='A', ta=[{'kind': 'sma', 'length': 50}, {'kind': 'sma', 'length': 200}], description='', created='Sunday January 23, 2022, NYSE: 11:36:06, Local: 15:36:06 PST, Day 23/365 (6.00%)')

In [16]:

watch.load("IWM")

Out [16]:

[i] Loaded IWM[D]: IWM_D.csv

	open	high	low	close	volume	dividends	stock splits	SMA_50	SMA_200
date
2000-05-26	34.332574	34.473957	34.167627	34.473957	74800	0.0	0.0	NaN	NaN
2000-05-30	34.968787	35.746395	34.968787	35.746395	57600	0.0	0.0	NaN	NaN
2000-05-31	35.864233	36.335510	35.864233	35.876015	36000	0.0	0.0	NaN	NaN
2000-06-01	36.612390	36.688972	36.612390	36.688972	7000	0.0	0.0	NaN	NaN
2000-06-02	38.350224	38.597645	38.350224	38.597645	29400	0.0	0.0	NaN	NaN
...	...	...	...	...	...	...	...	...	...
2022-01-14	211.990005	214.399994	210.580002	214.309998	42004100	0.0	0.0	224.026839	222.766286
2022-01-18	212.199997	212.440002	207.529999	207.830002	49434400	0.0	0.0	223.426888	222.689057
2022-01-19	208.809998	209.460007	204.380005	204.449997	46209100	0.0	0.0	222.694334	222.598203
2022-01-20	205.380005	208.990005	200.330002	200.750000	50466000	0.0	0.0	221.872826	222.506812
2022-01-21	199.759995	203.020004	196.990005	196.990005	84702700	0.0	0.0	221.004233	222.387342

5449 rows × 9 columns

Running Simple Study B

In [17]:

# Load custom_b into Watchlist and verify
watch.study = custom_b
watch.study

Out [17]:

Study(name='B', ta=[{'kind': 'ema', 'length': 8}, {'kind': 'ema', 'length': 21}, {'kind': 'log_return', 'cumulative': True}, {'kind': 'rsi'}, {'kind': 'supertrend'}], description='', created='Sunday January 23, 2022, NYSE: 11:36:06, Local: 15:36:06 PST, Day 23/365 (6.00%)')

In [18]:

watch.load("IWM")

Out [18]:

[i] Loaded IWM[D]: IWM_D.csv

	open	high	low	close	volume	dividends	stock splits	EMA_8	EMA_21	CUMLOGRET_1	RSI_14	SUPERT_7_3.0	SUPERTd_7_3.0	SUPERTl_7_3.0	SUPERTs_7_3.0
date
2000-05-26	34.332574	34.473957	34.167627	34.473957	74800	0.0	0.0	NaN	NaN	0.000000	NaN	0.000000	1	NaN	NaN
2000-05-30	34.968787	35.746395	34.968787	35.746395	57600	0.0	0.0	NaN	NaN	0.036245	NaN	NaN	1	NaN	NaN
2000-05-31	35.864233	36.335510	35.864233	35.876015	36000	0.0	0.0	NaN	NaN	0.039865	NaN	NaN	1	NaN	NaN
2000-06-01	36.612390	36.688972	36.612390	36.688972	7000	0.0	0.0	NaN	NaN	0.062272	NaN	NaN	1	NaN	NaN
2000-06-02	38.350224	38.597645	38.350224	38.597645	29400	0.0	0.0	NaN	NaN	0.112987	NaN	NaN	1	NaN	NaN
...	...	...	...	...	...	...	...	...	...	...	...	...	...	...	...
2022-01-14	211.990005	214.399994	210.580002	214.309998	42004100	0.0	0.0	216.510909	218.695431	1.827219	40.838119	226.149780	-1	NaN	226.149780
2022-01-18	212.199997	212.440002	207.529999	207.830002	49434400	0.0	0.0	214.581819	217.707665	1.796516	33.120092	224.599099	-1	NaN	224.599099
2022-01-19	208.809998	209.460007	204.380005	204.449997	46209100	0.0	0.0	212.330303	216.502422	1.780119	29.941457	221.623519	-1	NaN	221.623519
2022-01-20	205.380005	208.990005	200.330002	200.750000	50466000	0.0	0.0	209.756902	215.070384	1.761856	26.898184	220.974446	-1	NaN	220.974446
2022-01-21	199.759995	203.020004	196.990005	196.990005	84702700	0.0	0.0	206.919814	213.426713	1.742949	24.205683	216.573097	-1	NaN	216.573097

5449 rows × 15 columns

Running Bad Study. (Misspelled indicator)

In [19]:

# Load custom_run_failure into Watchlist and verify
watch.study = custom_run_failure
watch.study

Out [19]:

Study(name='Runtime Failure', ta=[{'kind': 'peret_return'}], description='', created='Sunday January 23, 2022, NYSE: 11:36:06, Local: 15:36:06 PST, Day 23/365 (6.00%)')

In [20]:

try:
    iwm = watch.load("IWM")
except AttributeError as error:
    print(f"[X] Oops! {error}")

[i] Loaded IWM[D]: IWM_D.csv
[X] Oops! 'AnalysisIndicators' object has no attribute 'peret_return'

In [ ]:

Indicator Composition/Chaining

When you need an indicator to depend on the value of a prior indicator
Utilitze prefix or suffix to help identify unique columns or avoid column name clashes.

Volume MAs and MA chains

In [21]:

# Set EMA's and SMA's 'close' to 'volume' to create Volume MAs, prefix 'volume' MAs with 'VOLUME' so easy to identify the column
# Take a price EMA and apply LINREG from EMA's output
volmas_price_ma_chain = [
    {"kind":"ema", "close": "volume", "length": 10, "prefix": "VOLUME"},
    {"kind":"sma", "close": "volume", "length": 20, "prefix": "VOLUME"},
    {"kind":"ema", "length": 5},
    {"kind":"linreg", "close": "EMA_5", "length": 8, "prefix": "EMA_5"},
]
vp_ma_chain_ta = ta.Study("Volume MAs and Price MA chain", volmas_price_ma_chain)
vp_ma_chain_ta

Out [21]:

Study(name='Volume MAs and Price MA chain', ta=[{'kind': 'ema', 'close': 'volume', 'length': 10, 'prefix': 'VOLUME'}, {'kind': 'sma', 'close': 'volume', 'length': 20, 'prefix': 'VOLUME'}, {'kind': 'ema', 'length': 5}, {'kind': 'linreg', 'close': 'EMA_5', 'length': 8, 'prefix': 'EMA_5'}], description='', created='Sunday January 23, 2022, NYSE: 11:36:06, Local: 15:36:06 PST, Day 23/365 (6.00%)')

In [22]:

# Update the Watchlist
watch.study = vp_ma_chain_ta
watch.study.name

Out [22]:

'Volume MAs and Price MA chain'

In [23]:

spy = watch.load("SPY")
spy

Out [23]:

[i] Loaded SPY[D]: SPY_D.csv

	open	high	low	close	volume	dividends	stock splits	VOLUME_EMA_10	VOLUME_SMA_20	EMA_5	EMA_5_LR_8
date
1993-01-29	25.645575	25.645575	25.517985	25.627348	1003200	0.0	0	NaN	NaN	NaN	NaN
1993-02-01	25.645562	25.809607	25.645562	25.809607	480500	0.0	0	NaN	NaN	NaN	NaN
1993-02-02	25.791391	25.882527	25.736710	25.864300	201300	0.0	0	NaN	NaN	NaN	NaN
1993-02-03	25.900745	26.155924	25.882517	26.137697	529400	0.0	0	NaN	NaN	NaN	NaN
1993-02-04	26.228839	26.301748	25.937205	26.247066	531500	0.0	0	NaN	NaN	25.937204	NaN
...	...	...	...	...	...	...	...	...	...	...	...
2022-01-14	461.190002	465.089996	459.899994	464.720001	95849600	0.0	0	8.524857e+07	78776895.0	466.839058	466.795420
2022-01-18	459.739990	459.959991	455.309998	456.489990	109709100	0.0	0	8.969594e+07	77486770.0	463.389369	465.217500
2022-01-19	458.130005	459.609985	451.459991	451.750000	109357600	0.0	0	9.327079e+07	77597910.0	459.509579	462.402221
2022-01-20	453.750000	458.739990	444.500000	446.750000	122379700	0.0	0	9.856332e+07	80226580.0	455.256386	458.363146
2022-01-21	445.559998	448.059998	437.950012	437.980011	201913500	0.0	0	1.173543e+08	87377745.0	449.497594	452.779043

7299 rows × 11 columns

In [ ]:

MACD BBANDS

In [24]:

# MACD is the initial indicator that BBANDS depends on.
# Set BBANDS's 'close' to MACD's main signal, in this case 'MACD_12_26_9' and add a prefix (or suffix) so it's easier to identify
macd_bands_ta = [
    {"kind":"macd"},
    {"kind":"bbands", "close": "MACD_12_26_9", "length": 20, "ddof": 0, "prefix": "MACD"}
]
macd_bands_ta = ta.Study("MACD BBands", macd_bands_ta, f"BBANDS_{macd_bands_ta[1]['length']} applied to MACD")
macd_bands_ta

Out [24]:

Study(name='MACD BBands', ta=[{'kind': 'macd'}, {'kind': 'bbands', 'close': 'MACD_12_26_9', 'length': 20, 'ddof': 0, 'prefix': 'MACD'}], description='BBANDS_20 applied to MACD', created='Sunday January 23, 2022, NYSE: 11:36:06, Local: 15:36:06 PST, Day 23/365 (6.00%)')

In [25]:

# Update the Watchlist
watch.study = macd_bands_ta
watch.study.name

Out [25]:

'MACD BBands'

In [26]:

spy = watch.load("SPY")
spy

Out [26]:

[i] Loaded SPY[D]: SPY_D.csv

	open	high	low	close	volume	dividends	stock splits	MACD_12_26_9	MACDh_12_26_9	MACDs_12_26_9	MACD_BBL_20_2.0	MACD_BBM_20_2.0	MACD_BBU_20_2.0	MACD_BBB_20_2.0	MACD_BBP_20_2.0
date
1993-01-29	25.645575	25.645575	25.517985	25.627348	1003200	0.0	0	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN
1993-02-01	25.645562	25.809607	25.645562	25.809607	480500	0.0	0	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN
1993-02-02	25.791391	25.882527	25.736710	25.864300	201300	0.0	0	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN
1993-02-03	25.900745	26.155924	25.882517	26.137697	529400	0.0	0	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN
1993-02-04	26.228839	26.301748	25.937205	26.247066	531500	0.0	0	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN
...	...	...	...	...	...	...	...	...	...	...	...	...	...	...	...
2022-01-14	461.190002	465.089996	459.899994	464.720001	95849600	0.0	0	0.667763	-1.252266	1.920029	-0.122515	2.439251	5.001017	210.045326	0.154245
2022-01-18	459.739990	459.959991	455.309998	456.489990	109709100	0.0	0	-0.342005	-1.809627	1.467622	-0.465015	2.346428	5.157870	239.635954	0.021877
2022-01-19	458.130005	459.609985	451.459991	451.750000	109357600	0.0	0	-1.507355	-2.379982	0.872627	-0.976476	2.234080	5.444636	287.416349	-0.082677
2022-01-20	453.750000	458.739990	444.500000	446.750000	122379700	0.0	0	-2.802061	-2.939750	0.137689	-1.794548	2.055545	5.905638	374.605587	-0.130843
2022-01-21	445.559998	448.059998	437.950012	437.980011	201913500	0.0	0	-4.484100	-3.697431	-0.786669	-3.011202	1.773740	6.558682	539.531303	-0.153910

7299 rows × 15 columns

In [ ]:

Comprehensive Study

MACD and RSI Momentum with BBANDS and SMAs and Cumulative Log Returns

In [27]:

momo_bands_sma_ta = [
    {"kind":"sma", "length": 50},
    {"kind":"sma", "length": 200},
    {"kind":"bbands", "length": 20, "ddof": 0},
    {"kind":"macd"},
    {"kind":"rsi"},
    {"kind":"log_return", "cumulative": True},
    {"kind":"sma", "close": "CUMLOGRET_1", "length": 5, "suffix": "CUMLOGRET"},
]
momo_bands_sma_Study = ta.Study(
    "Momo, Bands and SMAs and Cumulative Log Returns", # name
    momo_bands_sma_ta, # ta
    "MACD and RSI Momo with BBANDS and SMAs 50 & 200 and Cumulative Log Returns" # description
)
momo_bands_sma_Study

Out [27]:

Study(name='Momo, Bands and SMAs and Cumulative Log Returns', ta=[{'kind': 'sma', 'length': 50}, {'kind': 'sma', 'length': 200}, {'kind': 'bbands', 'length': 20, 'ddof': 0}, {'kind': 'macd'}, {'kind': 'rsi'}, {'kind': 'log_return', 'cumulative': True}, {'kind': 'sma', 'close': 'CUMLOGRET_1', 'length': 5, 'suffix': 'CUMLOGRET'}], description='MACD and RSI Momo with BBANDS and SMAs 50 & 200 and Cumulative Log Returns', created='Sunday January 23, 2022, NYSE: 11:36:06, Local: 15:36:06 PST, Day 23/365 (6.00%)')

In [28]:

# Update the Watchlist
watch.study = momo_bands_sma_Study
watch.study.name

Out [28]:

'Momo, Bands and SMAs and Cumulative Log Returns'

In [29]:

spy = watch.load("SPY")
# Apply constants to the DataFrame for indicators
spy.ta.constants(True, [0, 30, 70])
spy.tail()

Out [29]:

[i] Loaded SPY[D]: SPY_D.csv

	open	high	low	close	volume	dividends	stock splits	SMA_50	SMA_200	BBL_20_2.0	...	BBP_20_2.0	MACD_12_26_9	MACDh_12_26_9	MACDs_12_26_9	RSI_14	CUMLOGRET_1	SMA_5_CUMLOGRET	0	30	70
date
2022-01-14	461.190002	465.089996	459.899994	464.720001	95849600	0.0	0	465.989207	438.476059	456.838988	...	0.308635	0.667763	-1.252266	1.920029	46.269081	2.897775	2.902879	0	30	70
2022-01-18	459.739990	459.959991	455.309998	456.489990	109709100	0.0	0	465.813499	438.746869	456.082958	...	0.015239	-0.342005	-1.809627	1.467622	38.451767	2.879907	2.898966	0	30	70
2022-01-19	458.130005	459.609985	451.459991	451.750000	109357600	0.0	0	465.510704	438.995167	455.169080	...	-0.121184	-1.507355	-2.379982	0.872627	34.804529	2.869469	2.891151	0	30	70
2022-01-20	453.750000	458.739990	444.500000	446.750000	122379700	0.0	0	465.099938	439.216139	451.428136	...	-0.137331	-2.802061	-2.939750	0.137689	31.419064	2.858339	2.880571	0	30	70
2022-01-21	445.559998	448.059998	437.950012	437.980011	201913500	0.0	0	464.544664	439.383706	445.365560	...	-0.170887	-4.484100	-3.697431	-0.786669	26.542270	2.838513	2.868801	0	30	70

5 rows × 23 columns

In [ ]:

Additional Study Options

The params keyword takes a tuple as a shorthand to the parameter arguments in order.

Note: If the indicator arguments change, so will results. Breaking Changes will always be posted on the README.

The col_numbers keyword takes a tuple specifying which column to return if the result is a DataFrame.

In [30]:

params_ta = [
    {"kind":"ema", "params": (10,)},
    # params sets MACD's keyword arguments: fast=9, slow=19, signal=10
    # and returning the 2nd column: histogram
    {"kind":"macd", "params": (9, 19, 10), "col_numbers": (1,)},
    # Selects the Lower and Upper Bands and renames them LB and UB, ignoring the MB
    {"kind":"bbands", "col_numbers": (0,2), "col_names": ("LB", "UB")},
    {"kind":"log_return", "params": (5, False)},
]
params_ta_Study = ta.Study(
    "EMA, MACD History, Outter BBands, Log Returns", # name
    params_ta, # ta
    "EMA, MACD History, BBands(LB, UB), and Log Returns Study" # description
)
params_ta_Study

Out [30]:

Study(name='EMA, MACD History, Outter BBands, Log Returns', ta=[{'kind': 'ema', 'params': (10,)}, {'kind': 'macd', 'params': (9, 19, 10), 'col_numbers': (1,)}, {'kind': 'bbands', 'col_numbers': (0, 2), 'col_names': ('LB', 'UB')}, {'kind': 'log_return', 'params': (5, False)}], description='EMA, MACD History, BBands(LB, UB), and Log Returns Study', created='Sunday January 23, 2022, NYSE: 11:36:06, Local: 15:36:06 PST, Day 23/365 (6.00%)')

In [31]:

# Update the Watchlist
watch.study = params_ta_Study
watch.study.name

Out [31]:

'EMA, MACD History, Outter BBands, Log Returns'

In [32]:

spy = watch.load("SPY")
spy.tail()

Out [32]:

[i] Loaded SPY[D]: SPY_D.csv

	open	high	low	close	volume	dividends	stock splits	EMA_10	MACDh_9_19_10	LB	UB	LOGRET_5
date
2022-01-14	461.190002	465.089996	459.899994	464.720001	95849600	0.0	0	468.286877	-1.449058	461.651753	472.560246	-0.002944
2022-01-18	459.739990	459.959991	455.309998	456.489990	109709100	0.0	0	466.141989	-2.102017	455.062217	475.541774	-0.019567
2022-01-19	458.130005	459.609985	451.459991	451.750000	109357600	0.0	0	463.525264	-2.742434	448.133262	475.270730	-0.039072
2022-01-20	453.750000	458.739990	444.500000	446.750000	122379700	0.0	0	460.475216	-3.343681	442.732223	470.963773	-0.052901
2022-01-21	445.559998	448.059998	437.950012	437.980011	201913500	0.0	0	456.385178	-4.177700	433.536305	469.539696	-0.058853

In [ ]:

Disclaimer

All investments involve risk, and the past performance of a security, industry, sector, market, financial product, trading Study, or individual’s trading does not guarantee future results or returns. Investors are fully responsible for any investment decisions they make. Such decisions should be based solely on an evaluation of their financial circumstances, investment objectives, risk tolerance, and liquidity needs.
Any opinions, news, research, analyses, prices, or other information offered is provided as general market commentary, and does not constitute investment advice. I will not accept liability for any loss or damage, including without limitation any loss of profit, which may arise directly or indirectly from use of or reliance on such information.

218 KiB Raw Blame History Unescape Escape

Pandas TA (pandas_ta) Studies for Custom Technical Analysis

Topics

What is a Pandas TA Study?

Study Requirements:

Optional Requirements:

Things to note:

Builtin Examples

All

Common

Creating Studies

Simple Study A

Simple Study B

Bad Study. (Misspelled Indicator)

Study Management and Execution with Watchlist

Initialize AlphaVantage Data Source

Create Watchlist and set it's 'ds' to AlphaVantage

Info about the Watchlist. Note, the default Study is "All"

Help about Watchlist

Default Study is "Common"

Easy to swap Studies and run them

Running Simple Study A

Running Simple Study B

Running Bad Study. (Misspelled indicator)

Indicator Composition/Chaining

Volume MAs and MA chains

MACD BBANDS

Comprehensive Study

MACD and RSI Momentum with BBANDS and SMAs and Cumulative Log Returns

Additional Study Options

Disclaimer

218 KiB

Raw Blame History