How-to.md

Getting started

Set up

Check python

python --version

Check conda environment

conda list

Update conda or all packages

conda upgrade conda
conda upgrade --all

Install a new package

conda install numpy

Or an specific version

conda install numpy=1.10

Remove a package

conda remove numpy

Search for a package

Make use of wildcards

conda search *beautifulsoup*

Install Jupyter Notebook

conda install jupyter notebook

Launch Jupyter Notebook

jupyter notebook

Install Notebook Conda

This is helpful to manage environments from Jupyter Notebook

conda install nb_conda

Using Jupyter Notebook

Magic Keywords

Magic Keywords documentation

%matplotlib

Use it to set up matplotlib from your notebook. For example, you can use Inline backend, and render images for retina resolution:

%matplotlib inline
config InlineBackend.figure_format = 'retina'

%pdb

Use it for debugging. Type q on the prompt to leave the debugger.

%timeit

Use it for timing instructions in a cell.

Converting Notebooks

You can convert notebooks to multiple formats with nbconvert:

jupyter nbconvert --to html notebook.ipynb

You can convert them to slides and serve them:

jupyter nbconvert notebook.ipynb --to slides --post serve

Reading and loading data

Using numpy

numpy.loadtxt('data.csv', delimiter = ',')

Using pandas

pandas.read_csv('data.csv')

Assigning specific columns to variables

X = data.iloc[:,:-1]
y = data.iloc[:,-1]

Remove column from dataframe

data.drop('Header', axis = 1)

Combining pandas and numpy

data = numpy.asarray(pandas.read_csv('data.csv', header=None))

Linear Regression

Simple Linear Regression

from sklearn.linear_model import LinearRegression
model = LinearRegression()
model.fit(X, y)
model.predict(X_test)

Polynomial Regression

from sklearn.preprocessing import PolynomialFeatures
from sklearn.linear_model import LinearRegression

poly_feat = PolynomialFeatures(degree = 4)
X_poly = poly_feat.fit_transform(X)
poly_model = LinearRegression(fit_intercept = False).fit(X_poly, y)
model.predict(X_test)

Using pipeline

from sklearn.preprocessing import PolynomialFeatures
from sklearn.linear_model import LinearRegression
from sklearn.pipeline import make_pipeline

poly_model = make_pipeline(PolynomialFeatures(degree=4), LinearRegression())
poly_model.fit(X, y)
model.predict(X_test)

Linearization

L1 Linearization (Lasso)

from sklearn.linear_model import Lasso

lasso_reg = Lasso()
lasso_reg.fit(X,y)
reg_coef = lasso_reg.coef_

For L2 linearization, Ridge can be used

Feature Scaling

Standard Scale

from sklearn.linear_model import Lasso
from sklearn.preprocessing import StandardScaler

scaler = StandardScaler()

X_scaled = scaler.fit_transform(X)
lasso_reg = Lasso()
lasso_reg.fit(X_scaled, y)
reg_coef = lasso_reg.coef_