Enhanced-Hammett

This repository provides a python library with the implementation of the Enhanced-Hammett model [1]

It requires numpy, scipy and pandas to function, as well as matplotlib for the visualization of results

Tutorial

This section provides a small tutorial with examples on how to use this library.

First of all import the library and the example dataset with it.

import hammett as ham

Initial data

The data should be stored in a Pandas Dataframe structure, similar to the one provided as example

data = ham.get_experimental_data_Hudson_1962

A $\rho$ will be calculated for each columns and a $\sigma$ for each row. Missing values can be handled, but they should be represented by np.NaN. There should be at least two datapoints per column in order to get the corresponding $\rho$. The example reports kinetic constants for substituted thiols (rows) which react with different benzylbromides (columns).

Find parameters

The calc_param function will compute a the values of $\sigma$, $\rho$ and $\k_0$.

rho, sigma, k0, dicrho, dicsigma, dick0 = ham.calc_params(data)

Prediction

It is possible to build a new Dataframe to compare the initial data with the prediction by calling the following function:

prediction = ham.evaluate(data)

Visualization

To visualize the quality of the prediction you can use the plot_correlation function

ham.plot_correlation(data, prediction)