Exporting


ReLiS allows you to export different information from the project. To export the information, open the Classification phase of your project. Under the Result menu, select Export. Click on corresponding Update file to generate the latest version of the file you wish to export.

Classification

This is a CSV file of the classification table with all the data extracted. Each row correponds to a reference. The columns correspond to the categories listed in the classification form. Categories, for which multiple values are specified, are separated with a |.

Papers

This is a CSV file of all the references imported in the project. The columns are: the reference ID, the BibTeX key, the title, the URL of the reference, the preview of the reference in HTML format, the abstract, the year published, and the search strategy.

Papers (BibTeX)

This is a BibTeX file of all the references imported in the project.

Papers included (BibTeX)

This is a BibTeX file of all the references included after the last screening phase.

Papers excluded (BibTeX)

This is a BibTeX file of all the references excluded after the last screening phase.

Papers Excluded Screening(CSV)

This is a CSV file of all the references excluded after the last screening phase. The columns are: the reference ID, the BibTeX key, the title, the preview of the reference in HTML format, the search strategy, the user who excluded the reference, the exclusion criteria selected, and the note left by the user as a comment.

Papers Excluded Classification (CSV)

This is a CSV file of all the references excluded only during the classification phase. The columns are: the reference ID, the BibTeX key, the title, the preview of the reference in HTML format, the user who excluded the reference, the exclusion criteria selected, the note left by the user as a comment, and the search strategy.

R configuration script

This redirects you to the R export configuration page. See the Export to R section below for more information on that page. Clicking on Generate R redirects you to the Exports page. From there, you can download the R scripts which you can run using R to compute different statistics on the results of the classification.

This page allows you to specify the type of variable scale for each category. We support the following scales:

  • Continuous: For variables that are numeric and continuous in nature. This is typically for int and real simple categories.
  • Nominal: For categorical variables. This is typically for lists and dynamic lists, including if they have multiple values. But it can also be used for other categories, such as simple bool.
  • Text: For textual fields. If this scale type is assigned to a category, it will be excluded from all of the statistical tests and plots.

R Library

This is an R script file that is required for the R configuration script.


Export to R

You can export the results of the classification to R and run statistical analysis on them. To this end, you must download the following files exported from your project: Classification, R configuration script, and R Library.

The file to execute in R is the R configuration script. It contains a list of commented lines which consist of all possible statistical tests and plots available in the R library. Uncomment the lines plots/tests you wish to consider then run the script.

Note: You may need to change the path to the R library file based on your download locations. On line 1 of the R configuration script, make sure it states the proper name name and path to the R Library file you downloaded:

source("relis_r_lib_demo1_relis.R") # Replace this with the name of your imported library file
On line 7 of the R Library file, make sure it states the proper name and path to the classification file you downloaded:
relis_data <- read.csv("relis_classification_demo1_relis.csv", header = TRUE) # Replace this with the name of your imported data file

The R library supports the following plots and statistical tests. They are accessible by using the correpsonding variable. For example, statistics_vector[['Temperature']] outputs the descriptive statistics of the Temperature variable (this variable corresponds to a numerical category in the data extraction form.

Descriptive statistics

  • Descriptive frequency tables (for all nominal variables): use the desc_distr_vector variable
  • Bar plots (for all nominal variables): use the bar_plot_vector variable
  • Statistics (for all continuous variables): use the statistics_vector variable
  • Box plots (for all continuous variables): use the box_plot_vector variable
  • Violin plots (for all continuous variables): use the violin_plot_vector variable

Evolution statistics

  • Evolution frequency tables (for each nominal variables, with respect to the publication years of the references): use the evo_distr_vector variable
  • Evolution plots (for all nominal variables, with respect to the publication years of the references): use the evolution_plot_vector variable

Comparative statistics

  • Comparative frequency tables (for all nominal variables): use the comp_distr_vector variable
  • Stacked bar plots (for all nominal variables): use the stacked_bar_plot_vector variable
  • Grouped bar plots (for all nominal variables): use the grouped_bar_plot_vector variable
  • Bubble charts (for all nominal variables): use the bubble_chart_vector variable
  • Fisher's exact test (for all nominal variables): use the fisher_exact_test_vector variable
  • Shapiro Wilk's test (for all continuous variables): use the shapiro_wilk_test_vector variable
  • Pearson's correlation test (for all continuous variables): use the pearson_cor_test_vector variable
  • Spearman's correlation test (for all continuous variables): use the spearman_cor_test_vector variable

Export to Python

You can export the results of the classification to Python and run statistical analysis on them. When you generate (button "update file") the environment, you will then be able to download the files required to run the statistical analysis on your local machine.

The file to execute in R is the R configuration script. It contains a list of commented lines which consist of all possible statistical tests and plots available in the R library. Uncomment the lines plots/tests you wish to consider then run the script.

Note: Every file you need will be downloaded inside the ZIP file. You should have a .csv, containing all your data, and 2 python scripts, the kernel and the playground, that are needed to generate a proper statistical analysis environment. Make sure every file are inside the same folder when the scripts are executed.

The Python statistical environment supports the following plots and statistical tests. They are accessible by using the correpsonding variable. For example, evo_plots([Variables.cocoa_origin], True) outputs the evolutive statistics of the cocoa_origin variable (this variable corresponds to a nominal category in the data extraction form).

Descriptive statistics

  • Descriptive frequency tables (for all nominal variables): use the desc_frequency_table variable
  • Bar plots (for all nominal variables): use the desc_bar_plot variable
  • Statistics (for all continuous variables): use the desc_statistics variable
  • Box plots (for all continuous variables): use the desc_box_plot variable
  • Violin plots (for all continuous variables): use the desc_violin_plot variable

Evolution statistics

  • Evolution frequency tables (for each nominal variables, with respect to the publication years of the references): use the evo_frequency_table variable
  • Evolution plots (for all nominal variables, with respect to the publication years of the references): use the evo_plot variable

Comparative statistics

  • Comparative frequency tables (for all nominal variables): use the comp_frequency_table variable
  • Stacked bar plots (for all nominal variables): use the comp_stacked_bar_plot variable
  • Grouped bar plots (for all nominal variables): use the comp_grouped_bar_plot variable
  • Bubble charts (for all nominal variables): use the comp_bubble_chart variable
  • Chi Squared test (for all nominal variables): use the comp_chi_squared_test variable
  • Note: Originally this was "Fisher's exact test", but this test had some problems, so we replaced it with chi_squared for an alternative solution.

  • Shapiro Wilk's test (for all continuous variables): use the comp_shapiro_wilk_test variable
  • Pearson's correlation test (for all continuous variables): use the comp_pearson_cor_test variable
  • Spearman's correlation test (for all continuous variables): use the comp_spearman_cor_test variable