Data Analysis & Exploratory Data Analysis Using Python

Clarify independent versus dependent variables with real-world examples, using y = f(x) to show how factors like sleep and fertilizer affect weight, grade, and height.

Learn descriptive statistics in Python using NumPy and pandas, computing mean, median, mode, range, and standard deviation; examine percentiles, skewness, distribution shape, and data types (qualitative and quantitative).

Analyze boxplots built with Python to identify outliers using q1, q3, and the interquartile range, and explore invoice data by region and segment with Seaborn and Matplotlib.

Learn arithmetic, geometric, and harmonic means, and apply Python's statistics library to real-world data in finance, biology, and economics to compare growth and costs.

Explore how scatter plots reveal relationships between variables, quantify strength with correlation r and r-squared, compare linear and polynomial regression in Excel and Python to predict outcomes.

Explore how Pearson, Spearman, and Kendall's tau measure correlations in data analysis using Python, highlighting when to use each nonparametric method for monotonic or ordinal relationships.

Identify which factor—digital ad spend, traditional ad spend, or discount—most drives sales by comparing slopes from single-variable regressions in the three-round analysis.

Explore conditional scatterplots, bubble charts, and heatmaps with seaborn to analyze age, charges, BMI, and smoker status, and learn to use pair plots and lines of best fit.

Learn how to forecast future sales with time series methods by decomposing data into seasonality, trend, and residual, then model with ARIMA or Prophet to produce forecast ranges.

Explore sampling with representativeness and randomness, then apply hypothesis testing using p-values, significance level, and type I and II errors to compare null and alternate hypotheses.

Examine parametric tests for means, association, and distribution, including paired and independent t tests, ANOVA, Pearson correlation, linear regression, chi-square goodness of fit, z, and one-sample t tests.

Learn to perform t-tests, including one-sample and two-sample (paired and unpaired), and interpret t statistics and p-values with assumptions and practical examples.

Anova tests differences among means across groups, controls type I error, and uses the f statistic and p values to compare between and within variability in one- and two-factor designs.

Learn how t tests, ANOVA, and chi square tests drive hypothesis testing in data analysis, guiding EDA and business decisions before applying machine learning.

Differentiate parametric and non parametric tests by distribution assumptions and robustness to outliers; apply t tests or ANOVA when assumptions hold, else use chi square or Mann-Whitney.

Explore non-parametric tests organized by sample type and purpose, covering median tests (Wilcoxon), association (Spearman, Kendall), and distribution tests (Kolmogorov–Smirnov, chi-square, sign), with practical examples.

Learn how the nonparametric Mann-Whitney U test ranks data to compare two independent samples, an alternative to the independent t test when normality is not met. Compute U, interpret p-values.

Compare related samples with the Wilcoxon signed rank test, a non-parametric alternative to the paired t test, by computing differences, ranking differences, and assessing significance with the test statistic w.

Apply the Kruskal-Wallis test, a non-parametric method for comparing three groups when ANOVA assumptions fail. Use it to evaluate customer satisfaction scores across three strategies and interpret the p-value.

Use Mood's median test to compare medians across three strategies via a contingency table, chi-square, and p-value, illustrating median-based central-tendency analysis and Python implementation.

Apply the Friedman test, a non-parametric alternative to repeated-measures analysis of variance, to compare three or more related groups using rank-based data, compute the chi-square statistic, and interpret the p-value.

Apply the chi square test to evaluate relationships between two non-numeric variables using a contingency table of observed and expected frequencies, testing independence with the stats module from Scientific Python.

Identify and remove or replace null values with median for numeric data or mode for non-numeric data, and encode non-numeric features into numeric form to improve model accuracy.

Identify outliers via visual inspection or interquartile range, then apply cube root or log transformations to the entire data set to mitigate their impact and discuss distribution.

Learn about multi collinearity and detect it with the variance inflation factor (VIF); drop high VIF variables (cutoff 10) to improve regression, with logistic regression as an exception.

Examine imbalanced datasets in classification versus regression. In an insurance example, inflating the no class helps balance the data and improve learning.

Scale numeric data by subtracting the mean and dividing by the standard deviation to standardize features like insurance charges, age, BMI, and children, improving deep learning model accuracy.

Walk through an insurance data set to perform exploratory data analysis: load data, handle nulls, encode non-numeric features, assess multicollinearity with VIF, filter by correlation, and visualize distributions.

Explore exploratory data analysis with clip and sweetviz libraries using an insurance charges dataset to understand data, perform univariate and bivariate analysis, and address missing values, outliers, and imbalanced data.

Develop an interactive EDA web app using the stream lit library to share data insights with stakeholders. Provide tables, graphs, and visualizations in a web-based interface to meet end-user expectations.

Discover the infrastructure for building streamlined apps by installing Anaconda, using notepad or notepad++ for Python files, and installing streamlit with the Anaconda prompt along with numpy and pandas.

Create headers, a sub header, and an app title with success, warning, and error messages in a streamlit data analysis app. Upload a file to trigger backend processing by a machine learning model.

Learn how to access a folder-stored CSV file, read its data, and display the contents in a web app for data analysis and AI-focused projects.

Create a file uploader button using a file uploader widget in Streamlit to upload files for processing and display results.

Develop a web app for exploratory data analysis on a banking dataset, cleaning nulls with mean and mode and visualizing with histograms, frequency charts, and a correlation scatter plot.

Explore logistic regression and exploratory data analysis techniques to improve model accuracy in a banking loan scenario. Learn how logit, encoding methods, and multicollinearity affect accuracy and the confusion matrix.

Explore unsupervised learning with clustering and anomaly detection to uncover customer segments and fraud patterns, using k-means, kNN, and the elbow method.

Learn to analyze and manipulate data with Pandas, a Python library built on NumPy. Create and work with data frames for tabular data and perform exploratory data analysis.

Explore data analysis with pandas by inspecting a dataframe and computing descriptive statistics with describe. Learn to transpose, sort, select columns, copy, update values, and group data to reveal patterns.