A Demo web application for creating projects/organising Raman data and using self organising maps to perform multivariate Raman analysis. The application is a wrapper for the Kohonen library developed here
This code is currently deployed here:
alternatively you can follow the instructions below to setup the application locally.
- Install meteor (https://www.meteor.com/install)
- clone the repository
- Run the following commands from the repository directory in a terminal window:
meteor npm install
meteor
Visit http://localhost:3000
- Visit the home page and create an account to get started
- Click on
NEW PROJECT
to build a new Raman analysis project - Give the project a name and define a set of labels for separate groups of data you'd like to study (e.g. diseased, healthy).
- From the projects page, click
VIEW
. Click on one of your labels to expand the table view. ClickADD TRAINING DATA
to uopload Raman spectra. - Click on
FILE
and select a folder from your computer containing files that match your label (e.g. diseased). Your Raman spectra should be exported as tab-separated .txt files, with a single spectrum in each file. The first column should be wavenumber (x), and the second the Raman intensity (y). ClickSAVE
. - Repeat the previous step for each label.
- Click on
MACHINE LEARNING
from the project page an setup parameters for the self organising map. ClickNEW
to build a model with the desired settings. - Once the model is built, click
LOAD
to view the model any toggel between theCLUSTERING
,FEATURES
andCLASSIFICATION
tabs.
Fix SCSS imports:
cd imports
ln -s ../node_modules/materialize-css