Schottky data analysis

code for data-analysis

The Schottky spectroscopy experiments highly rely on the algorithms useful fr Schottky data analysis. The repo presented in this page arts as a library of the functions for data analysis, which includes common spectral estimator and analysis methods, such as numerical calculation of discrete prolate spheroidal sequence (DPSS), spectral density estimation of Schottky signal, periodogram with different windows, multitaper, and so on.

All the work may require the possession of basic digital signal processing knowledge at the graduate level to grasp the scientific rationale under the code. However, for the users only want to obtain a nice plot of the Schottky power spectrogram, applying the code according to some basic parameters is enough.

Code design

The code includes three scripts written in Python 3. It is just a library of data analysis, you need to import it in our own script or notebook like Jupyter.

The functional scripts:

• preprocessing.py: reading the header and loading the data (details)
  • the module extract and parse the called data file as default.
    • extract_wv() is applied for the data files (one acquisition includes a .wvh file as header and a .wvd file as raw data) from the R&S device.
    • extract_tiq() is applied for the data files (only a .tiq file for each acquisition) from the Tektronix device.
  • optional functions for the called data file are as following.
    • display() is for printing all the data acquisition parameters onto the screen.
    • load() is for loading the data into memory.
    • draw() is for plotting the signal in the in-phase and quadrature parts as a function of time.
    • diagnosis() is for checking the data by plotting them.
• processing.py: data analysis using several methods (details)
  • analysis methods: all methods provide two functions for results of 1d spectrum with windowed time series and 2d spectrogram with the whole acquisition time series.
    • fft_1d(), fft_2d(): FFT method
      computing the Fourier transform
    • spectrom(), spectrogram: FFT method
      computing the amplitude of the Fourier transform
    • periodogram_1d(), periodogram_2d(): FFT method
      estimating the power spectral density with FFT method
    • multitaper_1d(), multitaper_2d(): multitaper method
      estimating the power spectral density with multitaper method
    • adaptive_multitaper_1d(), adaptive_multitaper_2d(): adaptive multitaper method
      estimating the power spectral density with adaptive multitaper method
    • time_average_1d(), time_average_2d(): can apply all the methods
      estimating the power spectral density with time average
  • plot methods
    • plot_1d(): 1d spectrum
    • plot_2d(): 2d spectrogram
  • error analysis
    • confidence_band(): return the confience band of the spectral density esimation at a given confidence level.
• DPSS.py: generating the Discrete Prolate Speroidal Sequenece (DPSS), which is required for multitaper and adaptive multitaper methods.(details)
  • gen_sequences(), plot_sequences(): generating the DPSSs, and plotting them in the time domain, respectively.
  • gen_spectra(), plot_spectra(): Fourier transforming the DPSSs into the frequency domain, and plotting the result, respectively.

Usage

First of all is to ensure the followings are installed.

• Python 3
• Scipy, Numpy
• BLAS and LAPACK libraries (essential for matrix computation)
• PyFFTW (essential for FFT)
• Matplotlib (optional, only if visualization is needed)
• IPython and Jupyter (optional, only if presentation in a notebook is needed)

Next is to create a main Python script or a Jupyter notebook and import all the codes as modules.

from preprocessing import Preprocessing
from dpss import DPSS
from processing import Processing

Reference Wiki for more explanations about the class methods and arguments.