Analytics & Data Science for managers & humanitarians

Learn how socio-economic reality differs from technical systems, and apply data methods that account for dynamics, non-normal distributions, representative sampling, outliers, and model-driven analysis.

A model represents our assumptions about a real object and is not the object itself, with its elements, components, properties, and relationships, guiding description, measurement, interpretation, and prediction.

Explore how analytics and intuition together transform data into knowledge and conceptual models, enabling a holistic picture and the exchange of understanding with others.

Explore how regularities, patterns, trends, and tendencies relate to analytics and data science, and why the course uses regularities as the core term.

Explore analytics as modern methods for discovering hidden regularities in data and performing predictive analysis, while distinguishing data, statistics, metrics, and kpis from dashboards.

Explore the PwC maturity levels of corporate analytics, from data collection and reports to insights and predictive analytics, with emphasis on practical methods.

Compare data-analysis tools—from Excel to PSPP, SPSS, Statistica, and R—and learn how analytics rely on human thinking, not software.

Learn how statistical analysis helps managers and humanitarians make decisions by selecting relevant data and using descriptive and analytical methods to reveal real relations between variables and patterns, not chance.

learn how to define the general population and build a representative sample using random, stratified, and serial approaches, and how to determine the sample size.

Calculate sample size to obtain a representative sample from the general population using population size, margin of error, and confidence level, with online calculators.

Identify and describe variables as data indicators that vary across cases, such as revenue, income, city, gender, and height, and understand how each observation has its own variable values.

Identify three main scales of data—nominal, ordinal, and interval—and explore how each scale classifies objects, measures informativity, and enables different analysis methods.

Organize data arrays for analysis as a flat 2D table of cases (rows) and variables (columns), with numeric values and separate cells; avoid merged cells.

Learn how hypotheses function across fields—from marketing campaigns to investments and investigations—testing assumptions with data, accepting or rejecting them, and using probability to assess relationships.

Explain the null and alternative hypotheses (H0 and H1), outline Type I and Type II errors, and show how error probabilities and p-values guide general population conclusions.

An introduction to probability, p-values, and hypothesis testing, framing the null and alternative hypotheses, significance level alpha, and the decision rules for p<0.05 and p>0.05.

Explore probability concepts, null and alternative hypotheses, and how error probability determines statistical significance. Learn to evaluate whether sample findings generalize to the population using a 5% threshold.

Explore the normal distribution, the bell-shaped Gauss curve, and its rarity in socio-economic reality; learn that many data deviate from normality, requiring non-parametric measures and normality tests.

Use descriptive statistics to summarize data and analytical statistics reveal regularities for prediction. Compare samples to populations, track variables on nominal, ordinal, and interval scales, and test hypotheses with p<0.05.

Explore when to use Excel for descriptive statistics and why PSPP serves as a free, IBM SPSS analogue for practical analytics, especially for humanitarians.

Identify how Excel data arrays form 2d tables, with rows as cases or objects and columns as variables, showing each cell’s value for a specific case.

PSPP offers a free alternative to SPSS for analysis, with no expiration or artificial limits, supports importing data from spreadsheets, text files, and databases, and highlights cloud-free options like jasp-stats.org.

Explore data arrays in PSPP, mapping cases as rows. View data and variables, and define value labels and the measure scale as nominal, ordinal, or interval.

Import data to PSPP from a csv exported from Excel by following file menu steps, choosing all cases, naming variables, and setting the correct separator before saving the PSPP file.

Descriptive statistics compress large data into a single representative value using central tendency and variability measures such as mean, median, mode, min, max, range, skewness, kurtosis, and standard deviation.

Learn how frequency distributions describe counts and percentages, visualize them with pivot charts and histograms, and compute frequencies in Excel using recoding and the frequency function.

Explore mean, the most common descriptive statistic for interval data, its use in normal distributions, and how outliers and inappropriate contexts, like nominal scales or salaries, mislead estimates.

Explore mode and median as alternatives to the mean for nominal, ordinal, and skewed data, understand how they resist outliers, and practice calculating them in Excel.

Discover how to identify minimum and maximum values as the lowest and highest measurements, with practical Excel steps using min and max functions for any variable.

Explore percentiles, quartiles, and deciles in rank-ordered data, including the interquartile range and median, and learn to compute them in Excel for management salary review.

Explore variability, including range, standard deviation, and variance, to understand how data diverge from the average and detect outliers. Use Excel's analysis toolpack for practical variation assessment.

Discover how skewness and kurtosis describe distribution shape, symmetry, and tail weight, interpret positive versus negative skewness, and use Excel's data analysis outputs to inform data interpretation.

Identify how outliers distort data, distinguish input errors from real observations, and assess their impact using IQR, skewness, and multi-variable contexts.

Discover descriptive statistics, using frequency and measures of central tendency (mean, mode, median) plus dispersion (standard deviation, range, IQR) to compress data into representative values, contrasting with analytical statistics PSPP.

Learn to compute descriptive statistics in PSPP with the Analyze menu, using Frequencies, Descriptives, and Explore to generate frequency tables, statistics, histograms with normal curves, and percentiles.

Explore how analytical methods use probabilities, significance of differences, and relationships among multiple variables to classify cases and make predictions in socio-economical reality.

Explore how descriptive statistics summarize a data set by focusing on individual variables, while analytical statistics reveal hidden regularities, differences between objects, and make predictions using factors.

Explore descriptive statistics by variable and simultaneous analysis to reduce dimensions and classify objects. Build predictive models to test differences between groups and forecast future outcomes.

Identify relationships between variables, distinguish real regularities from randomness, and apply null and alternative hypotheses, with a significance level below 0.05 to gauge error risk.

Assess normal distribution to choose parametric or nonparametric criteria, considering sample size and central limit theorem, and test normality with Kolmogorov-Smirnov in PSPP or Excel.

Compare control and experimental groups across contexts—from mice trials to advertising—to determine if differences are significant, and apply parametric or non-parametric methods using Excel or PSPP.

Identify statistically significant differences between groups and assess the probability that findings are due to chance in the sample or general population.

Differentiate dependent (paired) samples from independent samples using before-and-after measurements and group comparisons. Apply these concepts to training data to assess significant improvements in metrics across groups.

Explore how crosstabs compare groups using pivot tables and chi-square significance, then recode reaction intensity in PSPP and assess results with lambda for nominal variables.

Compare independent samples using nonparametric methods, illustrated by AR versus classic education studying NumError in a non-normal distribution; apply Kruskal-Wallis and Median tests in PSPP and interpret significant results.

Compare paired (dependent) samples with independent ones to test before-and-after changes using non-parametric tests such as Wilcoxon and the sign test, and interpret significance at 0.05.

Compare groups to understand differences and their significance, using independent and paired samples, parametric and non-parametric criteria, with tests like t-Student, Wilcoxson, median and sign tests, significance less than 0.05.

Identify independent variables as the manipulated or predictor variables and the dependent variable as the observed outcome. Explain that correlation does not imply causality and model limitations apply.

Explore how socio-economic phenomena emerge from many interacting variables and how relationships are statistical, not strictly functional, with correlations and causation, and false correlations due to third variables.

Assess how likely a relationship found in a sample reflects the general population by testing the null hypothesis and evaluating statistical significance at a level below 0.05.

Explore how correlations reveal relationships between variables and measure strength with Pearson, Spearman, and Kendall coefficients. Understand linear and non-linear patterns, and beware false correlations and causation.

Explore regression and correlation as core tools of predictive analytics, using linear regression to predict loyalty from service quality and assortment while assessing model quality with R square and significance.

Learn how factor analysis reduces several hundred variables by grouping strongly correlated variables into factors, using eigenvalues, scree plots, and loadings to interpret and name underlying dimensions.

Explore reliability analysis, focusing on internal consistency and Cronbach's alpha, and learn to build a single-factor scale measuring loyalty and motivation from survey items.

Explore dependent and independent variables, distinguish correlation from causality, and review regression, factor analysis, and reliability analysis as tools for linear relationships, noting non-linear cases require different approaches.

Learn the classification task in analytics, the process of grouping objects by similarity or difference, and its role in machine learning, with examples: credit repayment, customer behavior, and spam detection.

Learn how binary logistic regression classifies cases into two groups using a probability-based approach, explains the model with Nagelkerke R square and probability calculation, and applies to business scenarios.

Cluster analysis groups objects with similar variable values into clusters, using nearest-neighbor merges and a chosen or auto-determined number of clusters, demonstrated with a PSPP k-means shopping data example.

Explore discriminant analysis, answer trees, and support vector machines, highlighting their use with nominal variables, comparisons to regression and clustering, and ROC visualization in binary classification.

Explore ensembles in data science that boost prediction accuracy by combining methods and data. Learn stacking, bagging (random forests), and boosting, including outlier handling and parallel computation.

Explore classification methods that predict group membership from variable values. Learn to apply logistic regression and cluster analysis, and recognize when ensembles improve forecasting, plus key terms in data management.

Explore how video, audio, text and other data types drive real-time analytics. Learn how online storage, powerful computing, and high-level languages empower regression, correlations, forecasting, and classification.

Explore how computing power, data warehouses, and networked devices enable big data, including unstructured data, and how analytics relates to this evolving data landscape.

Explore how big data and ensembles of analytical algorithms yield artificial intelligence, clarify its real-world nature, and connect machine learning with analytics and human-machine interaction.

Learn how machine learning uses analytical methods and algorithms to solve regression and classification tasks, with supervised and unsupervised approaches, ensembles, and neural networks.

Neural networks learn by adjusting weights through simple multilayer perceptrons to deep architectures, enabling forecasting and classification tasks in business, medicine, IT, and beyond.

Understand that AI is broad, machine learning is one stream, and neural networks are one method among many; use methods to find relationships, reduce dimensions, compare groups, and classify.