Visual analytics of high-dimensional data using a topology based approach. Regulus support both Morse and Morse-Smale complexes.
Regulus uses Bottle python server, which can installed via
pip install bottle
or conda install bottle -c conda-forge
.
Install the yarn package manager. e.g. conda install yarn -c conda-forge
.
- Run
yarn install
in the regulus root directory. It will locally install all the required Javascript packages. - Run
yarn build
to build the Javascript code. - During development, use
yarn watch
to continuously rebuild Regulus you save files.
Regulus works with high-dimensional data where the first d-dimensions are independent variables (dimensions) and the rest of dependent variables (measures).
- todo: [expand]: convert csv file to a regulus json file
- todo: [expand]: compute Morse- or Morse-Smale complexes
- todo: [expand]: compute various statistic such as linear regression and inverser regression
Regulus can sample the domain to create additional data points during the analysis process. There are two main approaches,
TBD
Currently, Regulus only support running a local cyclus for a specific scenario.
cd server; python server.py -d <path-to-data-directory>
You can also set an environment variable REGULUS_DATA_DIR
to point at the data directory.
Point your browser to localhost:8081
Regulus automatically save and reload the layout configuration.
localhost:8081/?noload
prevents loading of previous layoutlocalhost:8081/?nosave
prevents saving the layout during this runlocalhost:8081/?noload&nosave
prevents both