TaylorFit

TaylorFit is a multivariate polynomial regression application that fits data to a predictive model consisting of a polynomial function, such as

f(x, y) = a*x + b*y + c*xy + d*x*y^2

where x, y, and xy are terms selected by the user, and a, b, and c are coefficients determined using least squares regression.

On the Web

The application works entirely client-side in your browser, so there's no need to download or install anything and no need to create an account. In order to be reasonably efficient, TaylorFit utilizes specific JavaScript primitives and browser capabilities that emulate native execution.

Frameworks used

• KnockoutJS
• CoffeeScript
• PugJS
• Stylus

Installation

1. Ensure npm is installed
   • If npm --version prints version info, you're good
   • Else, install it at nodejs.org
2. Ensure yarn is installed
   • If yarn --version prints version info, you're good
   • Else, run npm install yarn to install
3. git clone or download a zip of the repo
4. Open the directory in a terminal
5. Run npm install
6. Run the server (by default hosted at http://localhost:8080/)
   • Run npm run debug to start the development server (includes hot reloading)
   • Run npm start to start the production server
7. Run npm run test to run the test suite

Export to gh-pages (the GitHub Pages site)

1. Run npm run build
2. Copy the build directory (cp -r build/ ../build/)
3. Run git checkout gh-pages
   1. If git won't let you checkout, use git status to see if you have any unsaved changes on the current branch
4. Paste the build directory files into the main directory (rm -r resources/ && mv ../build/* . && rm -r ../build/ [be careful you don't have another build directory in one level up from your clone])
   1. The first command deletes the old resources folder (since mv doesn't allow merges)
   2. The second command actually puts in the new build files
   3. The third command cleans up the copy of the build directory
5. Push up the code (git add -A && git commit -m "..." && git push)
6. Run git checkout master to return to the working branch