Smoothing

Smoothing¶

TAP data typically has some sort of noise within the outlet flux. This may be due to the heater, the voltage, or other external/environmental factors. To account for the noise without modifying the overall flux behavior, this package uses the Generalized Additive Model (GAM) to smooth the flux via splines. Below is an example of smoothing:

In [6]:

# importing tapsap, pandas, numpy and the data
import tapsap
import pandas as pd
import numpy as np
import plotly
import plotly.express as px
plotly.offline.init_notebook_mode()
argon_data = pd.read_csv('../tapsap/data/argon_100C_subset.csv')
times = argon_data['times'].values
flux = argon_data['pulse_75'].values

In [7]:

# apply smoothing to the flux
smoothed_flux = tapsap.smooth_flux_gam(flux)
df = pd.DataFrame({'raw':flux, 'smoothed':smoothed_flux})
tapsap.plot_tap(times, df, legend = True)

In [8]:

df_hist = pd.DataFrame()
df_hist['Smoothing Residuals'] = df['raw'].values - df['smoothed'].values
fig = px.histogram(df_hist, x='Smoothing Residuals', nbins=50, template='plotly_white')
fig.update_layout(
    font={
        'size': 30
    }
)
fig.update_layout(legend_title_text='', width = 1500, height = 500)
fig.show()

This method may also be applied to a Transient object through the use of smooth_flux. Note that the flux will be replaced by the smoothed values. Note that all smoothed values will be placed in a new data frame called smoothed_flux.

In [4]:

# read in the data and apply smoothing
data = tapsap.read_tdms('../tapsap/data/argon_100C.tdms')
transient_info = data.species_data['AMU_40_1']
transient_info.smooth_flux()
df = pd.DataFrame()
df['Non Smoothed'] = transient_info.flux.iloc[:,15].values
df['Smoothed'] = transient_info.smoothed_flux.iloc[:,15].values
tapsap.plot_tap(transient_info.times, df, legend=True)