Common financial technical indicators implemented in Pandas.
This is work in progress, bugs are expected and results of indicators might not be correct.
Dependencies:
- python (3+)
- pandas (0.18+)
TA class is very well documented and there should be no trouble
exploring it and using with your data. Each class method expects proper
ohlc data as input.
How to:
git clone https://github.com/peerchemist/finta && cd finta
sudo python setup.py install ## to install globally
or
pip install --user . ## to install locally (as user) which is prefered
from finta import TA
Prepare data to use with Finta:
finta expectes properly formated ohlc dataframe, with column names in lowercase ["open", "high", "low", close"] and ["volume"] for indicators that expect ohlcv input.
To prepare dataframe into ohlc format you can do something as following:
df.columns = ["date", 'close', 'volume'] ## standardize column names of your source
df.index = df["date"] ## set index on the date column, which is requirement to sort it by time periods
ohlc = df["close"].resample(24h).ohlc() ## select only price column, resample by time period and return daily ohlc (you can choose different time period)
ohlc() method aplied on the Series above will automatically format the dataframe in format expected by the library so resulting ohlc Series is ready to use.
Examples:
TA.SMA(ohlc, 42) ## will return Pandas Series object with Simple
moving average for 42 periods
TA.AO(ohlc) ## will return Pandas Series object with "Awesome oscillator" values
TA.OBV(ohlc) ## expects ["volume"] column as input
TA.BBANDS(ohlc) ## will return Series with Bollinger Bands columns [upper_bb, SMA, lower_bb, b_bandwith, percent_b]
TA.BBANDS(ohlc, TA.KAMA(ohlc, 20)) ## will return Series with calculated BBANDS values but will use KAMA instead of MA for calculation, other types of Moving Averages are allowed as well.
I welcome pull requests with new indicators or fixes for existing ones. Please submit only indicators that belong in public domain and are royalty free.
Some of the code is based from pandas_talib project but it is so radically changed that fork would not be a valid option. I urge the authors to merge with my code manually and continue developing on it.