t-SNE-Analysis

Custom collections of several tSNE-related plotting function.

Get Started

First, clone the repository and install all dependencies.

git clone https://github.com/marlenew27/t-SNE-Analysis.git
python -m pip install -r requirements.txt

If you have both Python 2 and Python 3 installed, please change the python above to python3.

Then, copy the tsnParams_template.json file and rename it into tsneParams.json. Edit it with your text editor.

Finally, run the code with python tsneMain.py

Usage

usage: tsneMain.py [-h] [-p PARAMETER]

Generate t-SNE models with data file and plot them

optional arguments:
  -h, --help            show this help message and exit
  -p PARAMETER, --parameter PARAMETER
                        path to JSON parameter file (default is
                        tsneParams.json in current directory)

JSON Parameter File

You may use -p or --parameter option to use another JSON parameter file.

Please use the tsneParams_template.json file to create your own JSON parameter file. The following code snippet is served as documentation, because comments are not allowed in standard JSON format. Please remove all comments (the part starting with // in each line) if you decide to use the following code snippet anyway.

{
    "file_path": "/file/path/to/excel/or/csv",  // the file name must end with .xlsx or .csv
    "groups_count": 2,
    "groups_info" : [                           // length of this array must be equal to groups_count
        {
            "group_name": "Group 1",            // optional
            "start_col": "A",                   // either Excel style or numeric string
            "end_col": "M",                     // either Excel style or numeric string
            "color": "red"                      // name or hex color
        },
        {
            "group_name": "Group 2",
            "start_col": "N",
            "end_col": "Z",
            "color": "green"
        }
    ],
    "region_col_index": "AA",                   // either Excel style or numeric string
    "fig_save_path": "fig_output.png",          // optional (relative path should be supported)
    "csv_save_path": "feature_weights.csv",     // path for csv file for PC weights (optional)
    "show_fig": false,                          // if true, pyplot.show() will be called
    "dimension": "2"                            // can be "2", "3", or "all" (must have the quotation marks)
}