EDA / Descriptive Statistics using Python (Part - 1)

Define the scope of application, set business objectives and constraints, and use input and output data to build a survival analytics model that reduces defaulters and increases profits.

Define the business success criteria and align with indicators, like reducing loan defaulters under 5% and machine learning accuracy above 85%, while defining crisp objectives and roi.

Explore business understanding and use cases, from fraud detection in credit card transactions to precision farming with drones, balancing fraud minimization, customer convenience, and cost constraints.

Explore data understanding by identifying data types, scales of measurement, and key terms, and compare primary and secondary data collection techniques.

Explore data understanding by measuring data, analyzing it to build models, and using what-if analysis to optimize sales under constraints for informed management decisions.

Compare continuous data and discrete data, showing how decimals define continuous measurements, with examples like time, money, height, and weight, and introducing numeric and categorical data.

Explore practical data understanding by comparing real-world examples to describe nominal, ordinal, interval, and ratio data, including absolute zero and subjective versus objective interpretations.

Explore the scale of measurement across nominal, ordinal, interval, and ratio data and learn which operations each type enables, highlighting ratio data as most suitable for analysis.

Differentiate quantitative data from qualitative data by showing numbers and continuous or count data. Identify qualitative data as descriptive or categorical and note its role in decision making.

Distinguish structured tabular data from unstructured media such as videos, images, audio, and text, and semi-structured forms like HTML, XML, and JSON, transforming raw data into structured formats.

Learn the essentials of data collection, including primary and secondary data sources, and the key terminology for variables, inputs, outputs, and structured data.

Understand secondary data sources and primary data sources, and learn to combine internal data with free open data, Google Maps, and drone analytics to generate richer, cost-efficient insights.

Discover primary data sources and how to augment bank data with external data like social media and IoT sensor data, while addressing data quality and privacy.

Understand how to translate business reality into survey data collection, identify root causes, and design construct-based questions to gauge training enrollment and price elasticity of demand.

Apply design of experiments to food promotions by testing expiry timing and distance; same-day expiry with near customers yields higher redemptions, revealing how expiry and distance affect consumer response.

Identify and mitigate data collection errors, including random and systematic errors, by checking device faults, wear, and human mismeasurement; establish standard operating procedures and gauge repeatability to ensure representative data.

Understand bias and fairness in data science, avoiding sensitive variables and biased data collection. Emphasize business understanding, proper data collection, and disciplined use of algorithms.

Introduce the CRISP ML(Q) data preparation framework and its phases, emphasizing business and data understanding, project charter, objectives, constraints, and success criteria, plus data types, data collection methods, and errors.

Learn the basic probability formula, P equals number of interested events divided by total events, and apply it to dice outcomes such as greater than three or less than four.

Define random variables by separating random and variable, using uppercase for the variable and lowercase for outcomes, and explore probability and distributions.

Explore how probability informs real-world decisions, from medical risk and legal cases to modeling random variables and probability distributions. Distinguish discrete versus continuous data with iPad sales examples.

Explore inferential statistics by understanding population, sampling frames, simple random sampling, and hypothesis testing to draw inferences from a data sample.

Review foundational concepts in probability, including discrete and continuous distributions, and inferential statistics with sampling and hypothesis testing, then preview exploratory data analysis for phase two.

Understand the normal distribution as a continuous probability distribution for a random variable, illustrated by heights, with histograms or density plots to assess its shape.

Explore the standard normal distribution, z-scores, and Six Sigma concepts; relate mean and standard deviation to a symmetric bell curve, and learn z-score standardization with mu and sigma.

Explore the first moment business decision by learning mean, median, and mode, and difference between population parameters and sample statistics. See how outliers affect mean and how median remains unaffected.

Explore measures of dispersion, or variation, the second moment business decision, using real-world profits across Malaysia and Singapore to assess forecast confidence and discuss control charts and outliers.

Explore the differences between percentiles, quantiles, and quartiles, and see how a box plot shows Q1 (first quartile), Q2 (median), and Q3 (third quartile) and their relation to percentiles.

Assess whether data are normally distributed using q-q plots alongside histograms and box plots, and learn about sample versus standardized quantiles and theoretical quantiles.

Learn how bivariate scatter plots reveal the direction and strength of relationships between two numerical variables, identify linear, non-linear, and exponential patterns, and gauge strength with the correlation r.

Learn how to install Python from python.org, choose the latest 3.10.7 release, on Windows, Linux, or macOS, and start using Python as a free, open-source tool.

Install and use the Anaconda distribution across Windows, macOS, and Linux, which ships with pre-installed libraries and is free for individuals but license-based for commercial use.

Explore Anaconda navigator and Spyder, and learn key Python libraries for data work. Practice reading csv with pandas, using read_csv, and understand zero-based indexing in data frames.

Learn to use Jupyter and Google Colab for Python data analysis, including running code, importing pandas, reading CSV files, uploading files, and leveraging GPUs or TPUs for faster practice.

Explore 1st and 2nd moment concepts for business decisions using Python, pandas for loading, and mean, median, mode, variance, and standard deviation with numpy, scipy, matplotlib, seaborn.

Examine the third moment and skewness in data using Python, linking mean, median, and standard deviation to histogram insights and real-world cases like Netflix analytics.

Explore the fourth moment and kurtosis using Python, linking excess kurtosis to normality, skewness, and business insights through retail and e-commerce examples.

Explore univariate data with bar plots and histograms in Python to convert raw numbers into meaningful insights, using bins, normal distribution checks, and simple examples like GMAT scores.

Explore univariate plots in Python with NumPy, SciPy, pandas, sklearn, statsmodels, matplotlib, and seaborn, including histograms, density plots, box plots, skewness, and kurtosis analysis.

Learn how to create and interpret univariate box plots in Python, identify outliers, and use the five-number summary (min, Q1, median, Q3, max) with histograms to assess distribution.

Explore univariate q-q plots with Python to assess normality, using SciPy stats, pandas to load data, and Matplotlib for visualization, with theoretical vs sample quantiles.

Explore creating a bivariate scatter plot in Python to relate waist circumference and adipose tissue, and interpret the correlation coefficient value of 0.81 and covariance.