Dfyjkt

I try to calculate the BBP ( Bollinger bands percent ) in python by this code. Howevery, my BBP function returns inf or -inf for bbp. Confusingly when I use some coin close price like ETH this function return the correct bbp number (not inf).

This is my python code:

import requests
import json 
import pandas as pd
import numpy as np
from talib import RSI, BBANDS

def BBP(price, close):
 up, mid, low = BBANDS(close, timeperiod=20, nbdevup=2, nbdevdn=2, matype=0)
 bbp = (price['close'] - low) / (up - low)
 print(up[-1])
 print(mid[-1])
 print(low[-1])
 print(bbp.iloc[-1])
 return bbp

r = requests.get('https://min-api.cryptocompare.com/data/histohour?fsym=SALT&tsym=BTC&limit=900&s=Binance&aggregate=5')
j = r.json()

price = pd.DataFrame(j['Data'])
price = price.sort_values(by='time', ascending=False)
price = price.iloc[::-1]
price = price.dropna()
close = price['close'].values

up, mid, low = BBANDS(close, timeperiod=20, nbdevup=2, nbdevdn=2, matype=0)

rsi = RSI(close, timeperiod=14)
bbp = BBP(price, close)

price.insert(loc=0, column='RSI',value=rsi)
price.insert(loc=0, column='BBP',value=bbp)

print(price.head(30))

If I use ETH instead of SALT in the request API that code works correctly but in other small prices coins the BBP function returns inf for the BBP column in the price data frame.

This is a sample of the return value for SALT:

 BBP RSI close high low open time 
0 NaN NaN 0.000069 0.000071 0.000068 0.000068 1534626000 
1 NaN NaN 0.000070 0.000070 0.000068 0.000069 1534644000 
2 NaN NaN 0.000072 0.000072 0.000068 0.000070 1534662000 
3 NaN NaN 0.000073 0.000073 0.000071 0.000072 1534680000 
4 NaN NaN 0.000074 0.000074 0.000072 0.000073 1534698000 
5 NaN NaN 0.000073 0.000074 0.000072 0.000074 1534716000 
6 NaN NaN 0.000073 0.000074 0.000072 0.000073 1534734000 
7 NaN NaN 0.000071 0.000073 0.000071 0.000073 1534752000 
8 NaN NaN 0.000072 0.000074 0.000070 0.000071 1534770000 
9 NaN NaN 0.000069 0.000072 0.000069 0.000072 1534788000 
10 NaN NaN 0.000070 0.000071 0.000068 0.000069 1534806000 
11 NaN NaN 0.000072 0.000072 0.000069 0.000070 1534824000 
12 NaN NaN 0.000070 0.000072 0.000070 0.000072 1534842000 
13 NaN NaN 0.000070 0.000070 0.000069 0.000070 1534860000 
14 NaN 56.138260 0.000071 0.000072 0.000069 0.000070 1534878000 
15 NaN 53.757682 0.000071 0.000073 0.000071 0.000071 1534896000 
16 NaN 56.547317 0.000072 0.000072 0.000070 0.000071 1534914000 
17 NaN 52.340624 0.000070 0.000072 0.000070 0.000072 1534932000 
18 NaN 42.426811 0.000067 0.000071 0.000067 0.000070 1534950000 
19 -inf 41.721667 0.000067 0.000067 0.000065 0.000067 1534968000 
20 -inf 41.087686 0.000066 0.000067 0.000066 0.000067 1534986000 
21 -inf 42.663976 0.000067 0.000067 0.000066 0.000066 1535004000 
22 -inf 46.241512 0.000068 0.000068 0.000066 0.000067 1535022000 
23 -inf 47.300220 0.000068 0.000069 0.000067 0.000068 1535040000 
24 -inf 47.984947 0.000068 0.000069 0.000067 0.000068 1535058000 
25 -inf 47.984947 0.000068 0.000069 0.000067 0.000068 1535076000 
26 -inf 50.590822 0.000069 0.000069 0.000068 0.000068 1535094000 
27 inf 56.805348 0.000071 0.000071 0.000068 0.000069 1535112000 
28 inf 57.658800 0.000071 0.000072 0.000069 0.000071 1535130000 
29 inf 63.418810 0.000073 0.000073 0.000070 0.000071 1535148000 

How can I fix this?

Thanks.

This is how I use BBANDS:

upperband, middleband, lowerband = talib.BBANDS(close, timeperiod=5, nbdevup=2, nbdevdn=2, matype=talib.MA_Type.T3)

I multiply close (and other prices) by 1000000 since I then rescale them. Maybe it is because the closing prices are too low.