Complete Data Science Training with Python for Data Analysis

Explore categorical, numerical, and ordinal data types used in statistical and machine learning analysis, with examples from surveys and measurements to guide technique selection.

Explore the three core Python data types: integers, floats, and strings, and learn to check types with the built-in type() function and convert between types using int and float.

Explore Python data science packages—NumPy for numeric operations, pandas for dataframes, Matplotlib and Seaborn for visualization, SciPy and statsmodels for statistics, scikit-learn for machine learning, and H2O for deep learning.

Explore numpy, a core python data science library providing fast multidimensional arrays and functions. Learn import conventions, convert lists to arrays, and apply arithmetic, solve linear equations, and compute statistics.

Create NumPy arrays from lists, explore rank-one to multi-dimensional arrays, and shape matrices for data science tasks. Inspect shapes, zeros matrices, and diagonal matrices to perform basic linear algebra operations.

Explore core numpy operations in Python, including indexing, slicing, negative indices, and multi-dimensional arrays, then practice copying, concatenating, and updating array elements for data analysis.

Explore matrix arithmetic and linear algebra with vectors and matrices, and see how Python performs addition, subtraction, scalar multiplication, dot product, matrix multiplication, inverse, and transpose.

Explore NumPy basic vector arithmetic by performing element-wise addition and subtraction on vectors X and Y, apply scalar addition, and compute element-wise multiplication, then extend to matrices.

Explore numpy matrix arithmetic with two-dimensional matrices, performing element-wise operations, scalar addition, and scalar multiplication, and standard matrix multiplication. Learn how to compute transpose and inverse, plus basic matrix-specific tasks.

Use numpy to solve linear systems by framing them as matrix equations, using the inverse of A and multiplying by B to obtain X and Y.

Explore NumPy for statistical operations, computing descriptive statistics like mean, median, standard deviation, and percentile on arrays, with axis-based summaries for columns and rows.

Explore Pandas data structures, including series and data frames, and learn to read data from external sources into data frames for practical data analysis and wrangling.

Learn to read CSV data using pandas, handling standard comma separated values and non-standard separators like semicolons or tabs, and extend to reading text and Excel formats.

Learn to read Excel files with pandas by loading the file path, listing sheet names, and loading a chosen sheet into a dataframe to preview data.

Learn to read JSON data in Python with pandas read_json, inspect a sample JSON featuring area, hectare criteria, description, image, and location, with 12 rows and seven columns.

Learn to read html data into Python data frames with pandas using read_html, scraping simple web pages and their tables, including handling multiple tables by index.

Read data from diverse sources, including structured and web-embedded formats, and learn to load Excel, JSON, and HTML data into pandas data frames for analysis in Python.

Leverage pandas to read diverse data and perform data wrangling tasks—cleaning, preprocessing, combining, reshaping, and cross tabulating—to prepare real-life datasets for analysis.

Identify missing values with isnull, remove or adjust them via drop or thresholding, and fill using zeros, forward fill, or backfill to ready the data for analysis.

Apply conditional data selection with pandas on a real endangered languages dataset from Kaggle, filtering languages with fewer than 5000 speakers and isolating key columns.

Learn to drop rows and columns in a data frame using pandas' drop function, removing the first row, the first two rows, and dropping columns by name (axis=1).

Master dataframe indexing and subsetting with pandas using iloc and loc to select rows and columns, apply conditional filters, and handle spaces in column names by renaming.

Cross tabulation analyzes language endangerment data by comparing two columns, country code and endangerment, using pandas crosstab with margins, then converts results to percentages for quick insights.

Reshape data with stack and melt to turn columns into rows and create tidy data with identifier variables and measured variables. Drop columns and remember reshaping is optional, not mandatory.

Master ranking and sorting data with pandas, including sort_index for series, sort_values for frames, and ranking data column-wise, such as by fit for service.

Learn how to concatenate data frames in Python with pandas, align common columns, handle missing values, and clean up column names for real-world data like Starbucks nutrition.

Merge data frames by a common country column, using left, right, and outer joins to combine GDP and global firepower data with metrics like manpower and railway coverage.

Master data wrangling with real-life datasets by handling missing values, indexing, grouping, and dropping columns, and applying ranking, sorting, cross tabulation, pivoting, and reshaping.

Utilize data visualization as a core data science tool to reveal patterns, trends, and correlations with bar plots, pie charts, histograms, and more in exploratory data analysis.

Explore the principles of data visualization and learn how data types: categorical, ordinal, continuous, and discrete, shape choosing histograms, bar plots, pie charts, line charts, box plots, and scatterplot matrices.

Visualize the distribution of continuous numerical variables with histograms, using univariate data and adjustable bins; explore Iris and GDP per capita datasets with matplotlib and seaborn.

Visualize distributions of continuous numerical variables with box plots, highlighting minimum, first quartile, median, third quartile, and maximum across iris species and tips data using seaborn.

Visualize relationships between two or more quantitative variables using scatter plots and scatter plot matrices with Pandas, Matplotlib, and Seaborn, including color by category and facets.

Visualize discrete data with bar plots and stacked bars using pandas and Seaborn, exploring country influence scores and Titanic survival by gender and class.

Learn how to visualize aggregated country-level influence with a pie chart, using sectors to show each country’s share and percentage values to compare USA, China, Japan, and others.

Summarize the end of section 6 by reinforcing data visualization principles, graph selection, and the use of Pandas, Matplotlib, and Seaborn for bar, pie, histograms, box plots, and scatter plots.

Explore what statistics are, including the collection, analysis, interpretation, and presentation of data, and how representative samples infer populations, while noting common misuses in polls.

Define the population and collect a representative random sample to ensure data quality and minimize bias. Distinguish observational from experimental studies, use stratified sampling, and apply p-values to test hypotheses.

Compute descriptive statistics on quantitative data using pandas. Load the iris dataset, convert to a data frame, and examine mean, standard deviation, min, max, percentiles, median, and interquartile range.

Group data by categories to summarize world university rankings by country and influence with describe and aggregate counts, then extend grouping to subdivision, annual, and year for mean and metrics.

Explore visualizing descriptive statistics with box plots using the iris dataset and seaborn to compare species, interpreting min, max, quartiles, median, IQR, and outliers.

Explore descriptive statistics by examining measures of center such as mean, median, and mode, compare outliers and skewed distributions, and quantify variation with standard deviation, standard error, and interquartile range.

Explore the normal distribution, a bell curve shape where mean, median, and mode are equal. Learn that parametric tests like linear regression rely on it, with 68/95/99.7 percent rules.

Evaluate if your data follow a normal distribution by inspecting a bell-shaped histogram and performing the Shapiro-Wilk test; recognize skewness and outliers that break normality and guide nonparametric choices.

Compute and interpret z-scores from the mean and standard deviation using the standard normal distribution, then apply z-score calculations to arrays and data frames in Python.

Explore the theory of confidence intervals, margin of error, and how sample statistics estimate population means, with focus on 95% intervals, z and t distributions, and sample size effects.

Calculate confidence intervals in Python, using the sample mean as point estimate, margin of error, and the critical value to capture the population parameter at a chosen confidence level.

Explore descriptive statistics, box plots, and data grouping to summarize quantitative data, and understand normal distribution and confidence intervals for analyzing relationships between variables.

Learn how to test claims about a population using null hypotheses, p-values, and alpha levels, and understand Type I and II errors and the test's power.

Compare means between two groups using parametric and non-parametric tests, including one- and two-sample t tests, Wilcoxon, Mann-Whitney U, and paired before-after analyses, with normality checks via the Shapiro test.

Apply one-way and two-way ANOVA to test differences among two or more groups, interpret p-values, and visualize results with box plots from iris and tooth growth data.

Explore the relationship between two quantitative variables using scatterplots and correlation to assess the strength of association between X and Y. Learn how regression modeling formalizes this relationship.

Explore relationships between two quantitative variables using correlation analysis, interpret scatterplots and correlation coefficients, and distinguish positive, negative, and no correlation before regression modeling.

Learn the theory of linear regression to model how Y changes with X, covering simple and multiple regression, the regression equation, and assessing significance with p-values and r-squared.

Identify whether linear regression assumptions are met—linear relation between X and Y, normally distributed and independent residuals with constant variance, and minimal multicollinearity.

Explore how to validate a linear regression model using residual normality tests (Jarque-Bera), QQ plots, Durbin-Watson, multicollinearity checks, heteroscedasticity tests (Breusch-Pagan), and influential points analysis on the iris dataset.

Explore polynomial regression to capture non-linear relationships by adding polynomial terms like x^2, x^3, and x^4, improving explained variance beyond linear models.

Explore generalized linear models to handle non-normal residuals, with Poisson and logistic regression for count and binary data, using exponential family and link functions, and implement them in R.

Explore logistic regression for binary outcomes using Python, modeling survival (0/1) with numeric and categorical predictors, and interpret odds, logit, and coefficients with Titanic data.

Concludes section eight by reviewing hypothesis testing, ANOVA, correlation and linear regression, introduces polynomial and generalized linear models, and logistic regression, and previews upcoming machine learning topics.

Compare statistics and machine learning: statistics formalizes variable relationships with equations and inference, while machine learning emphasizes predicting unseen data without explicit equations, guided by predictive accuracy.

Discover the basic theory of machine learning, contrast unsupervised and supervised classification, and see how data—not formal equations—drives prediction, clustering, and remote sensing analysis.

Explore k-means clustering, an unsupervised method that partitions data by assigning observations to the nearest centroid using Euclidean distance, iterating to minimize within-cluster variance and maximize between-cluster variance.

Apply k-means clustering to a real CSP glass dataset, encode seven glass types, and fit a seven-cluster model to predict categories from quantitative variables.

Explore hierarchical clustering, which forms nested clusters in a hierarchical tree, using agglomerative methods with distance measures like Euclidean, Manhattan, and Cosine.

Explore principal component analysis, an ordination and dimensionality reduction technique that converts correlated predictors into uncorrelated principal components, explaining maximum variation and enabling analysis with or without labels.

Learn to implement principal component analysis in Python, standardize numeric data, fit four components, and interpret variance explained for regression or classification.

Explore unsupervised learning concepts, including clustering, classification, and dimensionality reduction via principal component analysis, and evaluate methods with the Adjusted Rand Index and confusion matrices.