Linear Regression and Logistic Regression in Python

Install Python and the Anaconda distribution for data science, including Jupyter notebook and libraries, with a brief crash course and guidance for Mac or Windows and VS Code setup.

Celebrate reaching this milestone and joining the top 50% of learners. Maintain motivation to complete the course and use the Q&A, AI assistant, and certificate access.

Open Jupyter notebook using three methods: Anaconda Navigator, Anaconda Prompt, or Command Prompt, and navigate the default directory, change folders with cd, and understand ipynb notebooks in your browser.

Learn how to set up and navigate a Jupyter notebook in Anaconda, create and run Python code in cells, and switch between code, markdown, and raw formats using essential shortcuts.

Learn Python basics by performing arithmetic operations—addition, subtraction, multiplication, division, exponentiation, and modulus—while mastering operator precedence, variable assignment, and comparison operators.

Learn strings in Python basics, including quotes (single and double), variables, and print formatting. Discover string operations like length, replace, indexing, and slicing with positive and negative indices.

Explore Python data types with lists, tuples, and dictionaries, including creation, indexing, slicing, stepping, and nesting; compare mutable lists to immutable tuples and access dictionary values by keys.

Explore numpy, pandas, and seaborn basics, open a Jupyter notebook, create numpy arrays and matrices, inspect shapes and dtypes, and generate values with range, linspace, and random matrices in Python.

Explore the pandas library in Python to import csv data, inspect with head and describe, and index with iloc and loc for efficient data manipulation.

Visualize data with seaborn by plotting histograms and scatter plots, tweaking KDE visibility, and using pair plots on the iris dataset for sepal and petal measurements and species.

Identify the two main data types, qualitative (categorical) and quantitative, and their subtypes, nominal vs ordinal and discrete vs continuous, to determine appropriate analyses.

Explore the types of statistics, including descriptive and inferential statistics, and learn how measures of center and dispersion describe data, with graphs like bar charts and histograms.

Describe data with frequency distributions for qualitative and quantitative data, convert to bar charts and histograms, and interpret normal distribution, skewness, and symmetry.

Explore the four measures of center—mean, median, mode, and mid-range—and learn when to use population versus sample mean, compute medians, identify modes, and apply to qualitative data.

Explain measures of dispersion—range, standard deviation, and variance—and how outliers affect range; show how variance and standard deviation quantify spread from the mean.

Understand the business context to identify relevant variables and gather the data, then distinguish primary research from secondary research using examples like cart abandonment in online sales.

Identify data needs from research, request internal and external sources, and perform quality checks to prepare clean, well-defined data for analysis, using cart abandonment as an example.

explains assembling a property pricing dataset of 506 observations and 19 columns, defines price as the dependent variable, and outlines a data dictionary with a primary key to join sources.

Launch jupyter notebook via anaconda navigator, set your working directory, and import numpy, pandas, and seaborn to load a house price csv with headers, inspect with df.head and df.shape.

Explore univariate analysis of tabular data using descriptive statistics, mean, median, mode, range, quartiles, standard deviation, categorical counts, and the extended data dictionary (edd) to identify outliers and missing values.

Perform exploratory data analysis (EDD) on the dataset to inspect variables and missing values. Use scatterplots and seaborn joint plots to reveal price relationships and identify outliers and skewness.

Identify outliers using box plots, scatter plots, and histograms, and impute or cap them with percentile, sigma, or exponential smoothing methods to improve prediction accuracy.

Identify and treat outliers in Python using percentile-based capping for n_hot_rooms and rainfall, and transform crime rate with log or square root to improve linearity.

Learn how to handle missing values by imputing with zero, mean, median, or segment mean, and by using the most frequent category for categorical variables, guided by business context.

Impute missing values in Python by using df.info to identify gaps, then fill with the mean via fillna, updating the data frame.

explains seasonality in time-based data, shows how to remove it by applying a multiplication correction factor using monthly means, normalizing sales data for better model fit.

Explore bivariate analysis with scatterplots and correlation matrices to identify linear relationships and multicollinearity, then transform variables using log, exponential, or polynomial forms for improved linear regression.

Identify and remove non-informative variables, such as those with a single value or low fill rate, and use bivariate analysis to validate relevant relationships before regression modeling.

Convert categorical data into numeric indicators using dummy variables; regression requires numeric data and uses n minus one 0/1 dummies, where all zeros denote the baseline category.

Create dummy variables in Python to convert categorical airport and waterbody categories into numerical features, then drop redundant columns like airport_no and waterbody_none to illustrate full negative correlation.

Explore how correlation matrices reveal whether two variables move together, showing positive, negative, or near-zero relationships, and avoid multicollinearity by selecting variables with stronger links to the dependent variable.

Compute the correlation matrix with df dot corr to identify strong and weak relations, and remove redundant variables to avoid multicollinearity, such as removing park.

Explore linear regression as a simple supervised learning tool and its least squares method to fit a linear model, using house pricing data to quantify prediction accuracy and variable effects.

Explore simple linear regression with one predictor, estimate beta0 and beta1 via ordinary least squares, and minimize the residual sum of squares to predict Y.

learn to assess the accuracy of linear regression coefficients using standard errors, confidence intervals (including 95%), and hypothesis testing to determine if the number of rooms predicts house price.

Assess model accuracy by examining residual standard error and r-squared. Learn how adjusted r-squared accounts for predictors and how TSS and RSS relate to explained variance.

Learn simple linear regression in Python using statsmodels and sklearn, fit models with an intercept and slope, and interpret p-values, r-squared, and seaborn plots of the regression line.

Explore multiple linear regression with predictors, interpreting beta coefficients as the effect of each predictor on price while holding others fixed, and review model fit metrics like r-squared and p-values.

Explore how the f statistic assesses whether a regression model's predictors impact the response, adjusting for multiple variables, and interpret it with its p value before checking individual t values.

Build a multiple linear regression in Python with statsmodels and sklearn, creating X from all vars except price, adding a constant, and interpreting coefficients, p-values, and r-squared for key predictors.

Learn to split data into training and test sets, evaluate mean squared error, and compare training versus test error using validation set, leave-one-out, and k-fold cross-validation.

Explore the bias-variance tradeoff by comparing model flexibility, variance, and bias to minimize the test error, recognizing overfitting and irreducible error.

Split data into test and train sets with train_test_split, train a linear regression model, and evaluate performance on the test set using r2_score.

Explore three classification models—logistic regression, K nearest neighbours, and linear discriminant analysis—and predict sold within three months using a preprocessed dataset of 506 transactions.

Create a dataframe DF and load data with pd.read_csv, adjust Windows path slashes to forward slashes, and set header=0 for the first row.

Explore two regression questions: prediction and inferential, using house data to predict sale within three months and measure variable impact with classifiers.

Explore why linear regression fails for classification, including when the response has more than two levels, probability interpretation, and sensitivity to outliers; learn how logistic regression overcomes these issues.

Explore logistic regression for credit default, using the sigmoid function to map balance, income, and student to a 0–1 probability and estimate coefficients via maximum likelihood.

Learn to build a logistic regression model in python with sklearn and stats model, using price to predict sold, fitting with x and y, and examining coef and intercept.

Analyze a logistic regression with one predictor (price), interpret beta0 and beta1, compute probability of y=1, and assess significance using the p value to decide if price affects the response.

Extend logistic regression to multiple predictors using maximum likelihood to estimate betas and predict probabilities with a 0.5 decision boundary; for multi-class responses, consider linear discriminant analysis.

Learn to build a logistic regression model with multiple predictors in Python using sklearn and statsmodels. Prepare X and y, fit the model, interpret coefficients and intercept, and predict probabilities.

Explore how a confusion matrix evaluates model predictions by comparing true versus predicted values, highlighting type one error (false positive) and type two error (false negative) and threshold adjustments.

Learn to generate probability outputs with logistic regression via clf_lr.predict_proba, set and customize decision thresholds (0.5, 0.3), and evaluate predictions using confusion matrices to reveal true/false positives and negatives.

Assess classifier performance using confusion matrices, true/false positives and negatives, and metrics like precision, sensitivity, specificity, and ROC AUC.

Compute precision, recall, and ROC AUC scores for logistic regression and other models in Python using sklearn metrics, and compare AUC across models like LDA and KNN.

Split data into training and test sets to assess model accuracy on unseen data, using confusion matrices and test error, and explore validation, leave-one-out, and k-fold cross-validation.

Split the data into train and test sets with train_test_split, train a logistic regression model on the training set, and evaluate accuracy and the confusion matrix on the test set.

Celebrate finishing the course and earn your certificate of completion; if the certificate is greyed out, check missing lectures and mark them complete to unlock the download.