Reputation: 31
I am trying to calculate relative strength index (RSI) using pandas and can't seem to properly adapt a solution provided here. Why isn't this returning a RSI series?
import pandas_datareader.data as web
import datetime
start = datetime.datetime(2018, 2, 8)
end = datetime.datetime(2019, 2, 8)
stock = 'TNA'
price = web.DataReader(stock,'yahoo', start, end)
n = 14
def RSI(series):
delta = series.diff()
u = delta * 0
d = u.copy()
i_pos = delta > 0
i_neg = delta < 0
u[i_pos] = delta[i_pos]
d[i_neg] = delta[i_neg]
rs = moments.ewma(u, span=27) / moments.ewma(d, span=27)
return 100 - 100 / (1 + rs)
print(rsi(price, n))
Upvotes: 0
Views: 4648
Reputation: 2720
Here is a shot in the dark because you did not provide very much context.
pandas.stats.moment.ewma
is no longer supported in 0.23.0. Exponentially weighed windows are now achieved using pd.Series.ewm
. This returns exponentially-weighted-windows object window object that cannot be used in any sort of equation without providing a method for the rolling window. Here is a list of the available methods:
rs.agg rs.apply rs.count rs.exclusions rs.max rs.median rs.name rs.skew r.sum
rs.aggregate rs.corr rs.cov rs.kurt rs.mean rs.min rs.quantile rs.std rs.var
I assume you copied the function above from here, which did not even seem to answer the OP. If you wanted to do this analysis with the series price.Close
with span n
and compute the mean
of each exponentially weighted window:
import pandas_datareader.data as web
import datetime
import pandas as pd
ewma = pd.Series.ewm
start = datetime.datetime(2018, 2, 8)
end = datetime.datetime(2019, 2, 8)
stock = 'TNA'
price = web.DataReader(stock,'yahoo', start, end)
n = 14
def RSI(series,n):
delta = series.diff()
u = delta * 0
d = u.copy()
i_pos = delta > 0
i_neg = delta < 0
u[i_pos] = delta[i_pos]
d[i_neg] = delta[i_neg]
rs = ewma(u, span=n).mean() / ewma(d, span=n).mean()
return 100 - 100 / (1 + rs)
print(RSI(price.Close,n))
Upvotes: 1