Skip to main content

Visualization tools for the ASReview project

Project description

ASReview-visualization

This is a plotting/visualization supplemental package for the ASReview software. It is a fast way to create a visual impression of the ASReview with different dataset, models and model parameters.

Installation

The easiest way to install the visualization package is to use the command line:

pip install git+https://github.com/msdslab/ASReview-visualization.git

Basic usage

After installation of the visualization package, asreview should automatically detect it. Test this by:

asreview --help

It should list the 'plot' modus.

Log files that were created with the same ASReview settings can be put together/averaged by putting them in the same directory. Log files with different settings/datasets should be put in different directories to compare them. It is advised to put these log files in the same directory.

As an example consider the following directory structure, where we have two datasets, called ace and ptsd, each of which have 8 runs:

├── ace
│   ├── results_0.h5
│   ├── results_1.h5
│   ├── results_2.h5
│   ├── results_3.h5
│   ├── results_4.h5
│   ├── results_5.h5
│   ├── results_6.h5
│   └── results_7.h5
└── ptsd
    ├── results_0.h5
    ├── results_1.h5
    ├── results_2.h5
    ├── results_3.h5
    ├── results_4.h5
    ├── results_5.h5
    ├── results_6.h5
    └── results_7.h5

Then we can plot the results by:

asreview plot ace ptsd

By default, the values shown are expressed as percentages of the total number of papers. Use the -a or --absolute-values flags to have them expressed in absolute numbers:

asreview plot ace ptsd --absolute-values

Plot types

There are currently three plot types implemented: inclusions, discovery, limits. They can be individually selected with the -t or --type switch. Multiple plots can be made by using , as a separator:

asreview plot ace ptsd --type 'inclusions,discovery'

Inclusions

This figure shows the number/percentage of included papers found as a function of the number/percentage of papers reviewed. Initial included/excluded papers are subtracted so that the line always starts at (0,0).

The quicker the line goes to a 100%, the better the performance.

alt text

Discovery

This figure shows the distribution of the number of papers that have to be read before discovering each inclusion. Not every paper is equally hard to find.

The closer to the left, the better.

alt text

Limits

This figure shows how many papers need to be read with a given criterion. A criterion is expressed as "after reading y % of the papers, at most an average of z included papers have been not been seen by the reviewer, if he is using max sampling.". Here, y is shown on the y-axis, while three values of z are plotted as three different lines with the same color. The three values for z are 0.1, 0.5 and 2.0.

The quicker the lines touch the black (y=x) line, the better.

alt text

Project details


Download files

Download the file for your platform. If you're not sure which to choose, learn more about installing packages.

Files for asreview-visualization, version 0.1.4
Filename, size File type Python version Upload date Hashes
Filename, size asreview_visualization-0.1.4-py3-none-any.whl (11.7 kB) File type Wheel Python version py3 Upload date Hashes View hashes
Filename, size asreview-visualization-0.1.4.tar.gz (6.6 kB) File type Source Python version None Upload date Hashes View hashes

Supported by

Elastic Elastic Search Pingdom Pingdom Monitoring Google Google BigQuery Sentry Sentry Error logging AWS AWS Cloud computing DataDog DataDog Monitoring Fastly Fastly CDN DigiCert DigiCert EV certificate StatusPage StatusPage Status page