Currently, a few data cleaning tactics are employed. First, since there is no methane data past 2012, the VOC data
before 2012 is cut. We only look at ethane and acetylene in this circumstance, so other VOC columns are removed. Any
row with a NaN value is completely dropped. Since methane is sampled more frequently than the VOC's, methane values
from +/- three hours from any given VOC data point are averaged and then used to calculate the ratio of VOC / ch4 at
any given VOC datapoint timestep.
"""
# Import libraries
from fileInput import fileLoad
import pandas as pd
import numpy as np
import matplotlib.pyplot as plt
# Import Data Sets from Excel
nmhcData = fileLoad(r"C:\Users\ARL\Desktop\Python Code\Data\NMHC.xlsx")
methaneData = fileLoad(r"C:\Users\ARL\Desktop\Python Code\Data\Methane.xlsx")
# Cleaning Up Data
nmhcData = nmhcData[nmhcData['DecYear'] >
2012] # Only need years past 2012 in VOC Data
reqRows = ['DecYear', 'ethane',
'acetylene'] # only need date, ethane, and acetylene
nmhcData = nmhcData[reqRows] # just get required rows
methaneData = methaneData.dropna(
axis=0, how='any') # Remove NaN values, entire row is removed
nmhcData = nmhcData.dropna(axis=0, how='any')
# Define preliminary variables
years = np.linspace(2012, 2018, num=((2018 - 2012) + 1)) # Array of Years
numYears = np.size(years) # Total number of years
def ratioCalc():
import numpy as np
from fileInput import fileLoad
from isleapyear import isleapyear
# Import Data Sets
nmhcData = fileLoad(r"C:\Users\ARL\Desktop\Python Code\Data\NMHC.xlsx")
methaneData = fileLoad(
r"C:\Users\ARL\Desktop\Python Code\Data\Methane.xlsx")
numYears = np.linspace(2012, 2018,
num=((2018 - 2012) +
1)) # Total Number of Years in Dataset
nmhcDateAll = nmhcData.loc[:, 'DecYear'] # nmhc dates
ch4Date = methaneData.loc[:, 'DecYear'] # methane dates
hrs3 = 3 * 60 * 60 # three hours in seconds
# Preallocate Ratio Matrices
ethaneMethane = np.full((np.size(numYears), 1033),
np.nan) # Columns are for each year
aceMethane = np.full((np.size(numYears), 1033),
np.nan) # Rows are for the actual ratio values
datesFinal = np.full((np.size(numYears), 1033),
np.nan) # Dates for these ratio arrays
for i in numYears: # MAIN LOOP
# Date Variables for given year
nmhcDate = nmhcDateAll.loc[(nmhcDateAll >= i) & (
nmhcDateAll < (i + 1))].values # gathers current year
nmhcDate = 1 + (
(nmhcDate - i) *
(365 + isleapyear(i)) * 24 * 60 * 60) # convert to seconds
methaneDate = ch4Date.loc[(ch4Date >= i) & (ch4Date < (i + 1))].values
methaneDate = 1 + (methaneDate - i) * (365 +
isleapyear(i)) * 24 * 60 * 60
# Yearly compound values
ethane = nmhcData.loc[(nmhcDateAll >= i) & (nmhcDateAll < (i + 1)),
'ethane'].values
ace = nmhcData.loc[(nmhcDateAll >= i) & (nmhcDateAll < (i + 1)),
'acetylene'].values
methane = methaneData.loc[(ch4Date >= i) & (ch4Date < (i + 1)),
'MR'].values
# Create Ratio Vectors
for j, value in np.ndenumerate(ethane): # LOOP: Ethane values
high = nmhcDate[
j] + hrs3 # current Ethane timestep in seconds + 3 hours
low = nmhcDate[
j] - hrs3 # current ethane timestep in seconds - 3 hours
# Get the average of all methane values between high and low
methaneAverage = np.mean(methane[(methaneDate[:] <= high)
& (methaneDate[:] >= low)])
ethaneMethane[np.where(numYears == i),
j] = value / methaneAverage # Fills out matrix
for k, value in np.ndenumerate(ace): # LOOP: Acetylene Values
high = nmhcDate[k] + hrs3 # Same process as above
low = nmhcDate[k] - hrs3
methaneAverage = np.mean(methane[(methaneDate[:] <= high)
& (methaneDate[:] >= low)])
aceMethane[np.where(numYears == i), k] = value / methaneAverage
nmhcDate = (nmhcDate / 60 / 60 /
24) + (i - 2012) * (365 + isleapyear(i))
location = datesFinal > np.Inf
location[np.where(numYears == i), 0:np.size(nmhcDate)] = True
np.place(datesFinal, location, nmhcDate)
return ethaneMethane, aceMethane, datesFinal