Ecosyste.ms: Awesome

An open API service indexing awesome lists of open source software.

Awesome Lists | Featured Topics | Projects

https://github.com/5tefan/py-netcdf-timeseries-gui

A Python Qt5 GUI supporting exploratory analysis and visualization of time series based data from NetCDF files.
https://github.com/5tefan/py-netcdf-timeseries-gui

analysis gui netcdf python time-series

Last synced: 19 days ago
JSON representation

A Python Qt5 GUI supporting exploratory analysis and visualization of time series based data from NetCDF files.

Awesome Lists containing this project

README 

        
# Python NetCDF Time Series Plot GUI



The Python NetCDF Time Series GUI (pyntpg) is a tool to visualize and 

do analysis with series data in NetCDF format. 



## Overview 



The goal of the project is to streamline analysis and visualization of

NetCDF time series data. The program allows you to produce basic time 

series and scatter plots without writing any code. The integrated Ipython

console allows the user to perform any kind of data processing possible

in Python and display the results. 



The software package brings together many powerful third party components, making use of:



* PyQt4 Python bindings to the Qt UI library

* netcdf4-python interface to the netCDF C library

* matplotlib python visualization and plotting library

* Jupyter IPython console and kernel, qtconsole

 

One goal of continued ddevelopment on the interface is focusing on improving the 

modularity such that adapting to new features may only require rewriting

limited pieces of the interface. The panel_configurer for new plot styles

or the dataset_tabs for reading new data types. 



## Usage



A tutorial workflow is described below 



 1. Open NetCDF files using the "Add Files" button on the dataset tab, or by 

    selecting dataset -> Open files from the menu. Multiple files may be 

    selected which will then be concatenated into one underlying netcdf file

    object. Note that the files must have the same format, otherwise 

    concatenation will fail.

 2. When the import is successful, the progress bar will disappear and an 

    ncinfo like preview of the file will be displayed in the right hand text box.

    At this point, a variable with the same name your dataset tab ("dataset" is

    the default for the first tab) is available in the IPython console, accessible

    from Edit -> Open IPython Console. More datasets can be added by clicking the

    "+" tab, or from File -> New Dataset.

 3. In the plot tab below, customize the panel layout to your preference. We will

    opt for two side by side panels for demonstration. Drag the black bars to the

    nearest edge until the number disappears and you are left with two equally 

    sized panels labeled 0 and 1. 

 4. Below, under "y axis picker", configure the From field to be "dataset", and 

    choose a variable. To make the tutorial simple, hopefully the variable you 

    selected only depends on the time dimension, but if not, options to flatten 

    into a 1 dimensional array should appear. 

 5. Under "x axis picker", select the datetime radio button, then again configure

    the From field to be "dataset", and choose the variable representing time. If

    the units are configured correctly, the date range of your data should display

    in the start and end boxes. You may configure the start and end to constrain

    what is shown.

 6. Stroke color and Stroke style should be ok, feel free to change the color. The

    Stroke style indicators are exactly from matplotlib, '-' is a solid line, '*' 

    makes each point a star, etc. See matplotlib documentation for more. The Panel 

    destination box indicates which panel the data you just configured will be 

    drawn on. You can add the line to panels that are not currently on the layout,

    they will appear if you change their number or add the panel number it is 

    specified on back to they layout. Click "Add to Queue" and the line should 

    appear in the "Queue to plot". You can right click the queued line to change

    color, line style, or the panel it is attached to at any time. 

 7. Next, we will create a scatter plot to add to panel 1 by changing, under "x

    axis picker" the radio button to other. Select another variable to plot 

    a scatter plot against. I suggest selecting line style '.' for scatter plot.

    Select panel destination 1 and Add to Queue. 

 8. Finally, to display the plot, click the green "Create Plot" button at the 

    bottom. A toolbar within the window that should appear allows you to zoom,

    pan, save, and even configure the axes (eg. log, linear, min, max) as well

    as options to change the title and labels.

    

 



# TODO



* Opening frequency analysis wizards takes a while, especially for lots of data.

Ie, complexity of opening a wizard is at least O(n) in size of data, which should not

be the case! from some initial profiling,

`python -m cProfile -s cumtime pyntpg/main.py > profile.log`, and code inspection,

this appears to be due to the calls from signal\_picker, specifically the get\_values 

call is expensive. 

* ability to transport/compose wizard plots to plot tab queue

* catch errors coming out of the analysis wizard and appropriately intercept in preview\_result