CODEWITHSUNDEEP
pythonpython-3.xpandasdataframealgorithmic-trading
MoonBoi9001
MoonBoi9001

Reputation: 65

How can I simplify code that runs the same calculation with different parameters into different output variables?

I have a dataframe, df, that contains 1.4 million rows of data, where each row represents 1 minute of open, high, low and close prices for BTC from 2018 to 2020. I want to add the MACD (Popular trading indicator) to my df, but instead of only calculating the macd for the 1 minute time frame like this:

ShortEMA = df.Close.ewm(span=12, adjust=False).mean()
LongEMA = df.Close.ewm(span=26, adjust=False).mean()
MACD = ShortEMA - LongEMA
signal = MACD.ewm(span=9, adjust=False).mean()

df["MACD"] = MACD
df["Signal Line"] = signal

I want to calculate the MACD for each of the time frames 1 minute, 15 minutes, 30 minutes, 1 hour, ect...

I did this with the following code (which took ages):

MySet = [1, 5, 15, 30, 60, 240, 360, 720, 1440, 10080]

ShortEMA1 = df.Close.ewm(span=12 * MySet[0], adjust=False).mean()
LongEMA1 = df.Close.ewm(span=26 * MySet[0], adjust=False).mean()
MACD1 = ShortEMA1 - LongEMA1
signal1 = MACD.ewm(span=9 * MySet[0], adjust=False).mean()

ShortEMA5 = df.Close.ewm(span=12 * MySet[1], adjust=False).mean()
LongEMA5 = df.Close.ewm(span=26 * MySet[1], adjust=False).mean()
MACD5 = ShortEMA5 - LongEMA5
signal5 = MACD.ewm(span=9 * MySet[1], adjust=False).mean()

ShortEMA15 = df.Close.ewm(span=12 * MySet[2], adjust=False).mean()
LongEMA15 = df.Close.ewm(span=26 * MySet[2], adjust=False).mean()
MACD15 = ShortEMA15 - LongEMA15
signal15 = MACD.ewm(span=9 * MySet[2], adjust=False).mean()

ShortEMA30 = df.Close.ewm(span=12 * MySet[3], adjust=False).mean()
LongEMA30 = df.Close.ewm(span=26 * MySet[3], adjust=False).mean()
MACD30 = ShortEMA30 - LongEMA30
signal30 = MACD.ewm(span=9 * MySet[3], adjust=False).mean()

ShortEMA60 = df.Close.ewm(span=12 * MySet[4], adjust=False).mean()
LongEMA60 = df.Close.ewm(span=26 * MySet[4], adjust=False).mean()
MACD60 = ShortEMA60 - LongEMA60
signal60 = MACD.ewm(span=9 * MySet[4], adjust=False).mean()

ShortEMA240 = df.Close.ewm(span=12 * MySet[5], adjust=False).mean()
LongEMA240 = df.Close.ewm(span=26 * MySet[5], adjust=False).mean()
MACD240 = ShortEMA240 - LongEMA240
signal240 = MACD.ewm(span=9 * MySet[5], adjust=False).mean()

ShortEMA360 = df.Close.ewm(span=12 * MySet[6], adjust=False).mean()
LongEMA360 = df.Close.ewm(span=26 * MySet[6], adjust=False).mean()
MACD360 = ShortEMA360 - LongEMA360
signal360 = MACD.ewm(span=9 * MySet[6], adjust=False).mean()

ShortEMA720 = df.Close.ewm(span=12 * MySet[7], adjust=False).mean()
LongEMA720 = df.Close.ewm(span=26 * MySet[7], adjust=False).mean()
MACD720 = ShortEMA720 - LongEMA720
signal720 = MACD.ewm(span=9 * MySet[7], adjust=False).mean()

ShortEMA1440 = df.Close.ewm(span=12 * MySet[8], adjust=False).mean()
LongEMA1440 = df.Close.ewm(span=26 * MySet[8], adjust=False).mean()
MACD1440 = ShortEMA1440 - LongEMA1440
signal1440 = MACD.ewm(span=9 * MySet[8], adjust=False).mean()

ShortEMA10080 = df.Close.ewm(span=12 * MySet[9], adjust=False).mean()
LongEMA10080 = df.Close.ewm(span=26 * MySet[9], adjust=False).mean()
MACD10080 = ShortEMA10080 - LongEMA10080
signal10080 = MACD.ewm(span=9 * MySet[9], adjust=False).mean()


df["MACD1"] = MACD1
df["Signal Line1"] = signal1

df["MACD5"] = MACD1
df["Signal Line5"] = signal5

df["MACD15"] = MACD1
df["Signal Line15"] = signal15

df["MACD30"] = MACD1
df["Signal Line30"] = signal30

df["MACD60"] = MACD60
df["Signal Line60"] = signal60

df["MACD240"] = MACD240
df["Signal Line240"] = signal240

df["MACD360"] = MACD360
df["Signal Line360"] = signal360

df["MACD720"] = MACD720
df["Signal Line720"] = signal720

df["MACD1440"] = MACD1440
df["Signal Line1440"] = signal1440

df["MACD10080"] = MACD10080
df["Signal Line10080"] = signal10080

How can I simplify this whole process?

Upvotes: 0

Views: 91

Answers (2)

Henry Ecker
Henry Ecker

Reputation: 35646

If the only persistent output are the values stored in the DataFrame and the intermediate Series are not going to be used or re-used later it may be better to simply update the DataFrame and assign columns with an f-string in each iteration:

MySet = [1, 5, 15, 30, 60, 240, 360, 720, 1440, 10080]
for val in MySet:
    ShortEMA = df.Close.ewm(span=12 * val, adjust=False).mean()
    LongEMA = df.Close.ewm(span=26 * val, adjust=False).mean()
    df[f"MACD{val}"] = ShortEMA - LongEMA
    df[f"Signal Line{val}"] = df[f"MACD{val}"].ewm(span=9 * val, adjust=False).mean()

If values need to be accessed later they can be accessed through the DataFrame.

Upvotes: 2

Charles Duffy
Charles Duffy

Reputation: 295472

Instead of having Short1 and Short5 be separate variables, have there just be one Shorts dictionary, and have 1, 5, etc be keys. Thus:

MySet = [1, 5, 15, 30, 60, 240, 360, 720, 1440, 10080]
Shorts = {}
Longs = {}
MACDs = {}
Signals = {}

for val in MySet:
  Shorts[val] = df.Close.ewm(span=12 * val, adjust=False).mean()
  Longs[val] = df.Close.ewm(span=26 * val, adjust=False).mean()
  MACDs[val] = Shorts[val] - Longs[val]
  Signals[val] = MACDs[val].ewm(span=9 * val, adjust=False).mean()

  df[f'MACD{val}'] = MACDs[val]
  df[f'Signal Line{val}'] = Signals[val]

Upvotes: 2

Related Questions