Business Analytics

Versión en español Syllabus (PDF, in Spanish) Week-by-week program (PDF) Course podcast · El Dato con Contexto

Original title (in Spanish): Analítica para los negocios (06327-ECO). All materials are in Spanish — a full Spanish version of this page is available.

The course materials are published on this site as each week is released. The documents are built with Quarto and embed runnable R through webR, so students can execute every example directly in the browser, with no local installation required.

Course description

The course introduces the business-analytics process: how to turn raw data into information, knowledge and evidence that supports decisions in private and public organizations. It concentrates on five analytical tasks — summarizing, visualizing, clustering, classifying, and estimating regressions — implemented in R on business datasets. Generative AI runs through the whole semester as a working tool: students learn what LLMs are, how to write effective prompts, and how to direct AI coding agents (Claude Code, Cursor, VS Code) responsibly — the AI accelerates the work, but the student must understand and validate every step of the analysis.

How each week works

Each content week publishes three materials, in the order students use them:

1. Theory — studied before class: the written version of the week’s podcast or video, with runnable R examples. The podcast episodes are published on Spotify — El Dato con Contexto — and each one is linked in its week below. The weekly quiz is printed and answered at the start of class — a few minutes of questions about the material first, then the quiz, with no AI or other aids.
2. Guided practice — the in-class application, led step by step.
3. Workshop — the graded activity each student completes during the session and submits on Intu before class ends.

Learning outcomes

By the end of the course, students will be able to:

• Explain the core concepts of business analytics and identify the analytical task — summarize, visualize, cluster, classify, regress — that answers a given business question.
• Use R to import, explore, clean, transform, and describe data, and build clear visualizations with ggplot2.
• Diagnose data-quality problems and carry an exploratory data analysis from raw file to analysis-ready dataset with documented decisions.
• Train and evaluate machine-learning models in R — CART, random forests, the Lasso, and k-means — using train/test splits and cross-validation, comparing against a baseline with the right metrics.
• Interpret and communicate results to support an organizational decision.
• Use generative-AI tools — LLMs and coding agents — critically and responsibly: effective prompts, validated outputs, traceable analysis.

Schedule

Unit 1 — Course foundations

Week 1 — Presentación del curso y metodología. What business analytics is and the problems it solves; the week-by-week tour; the learning dynamic (theory before class, quiz, guided application); evaluation rules, the final project, and the AI policy — motivated by recent evidence on how AI use affects skill formation.

Theory Reading: Shen & Tamkin (2026)

Week 2 — Introducción a los LLMs. What a large language model is and why it matters for economics and business; capabilities, limits and hallucinations; the anatomy of an effective prompt (context, task, output format, constraints); reusable skills for recurring tasks.

Theory Podcast Guided practice Workshop

Week 3 — Fundamentos de R. The RStudio interface and the script-based workflow; R as a calculator; data types and special values; objects and assignment; vectors, matrices and data frames; functions, help, and packages; the Environment and projects with relative paths.

Theory Guided practice Workshop Video · R fundamentals

Week 4 — Manipulación y visualización de datos. The dplyr grammar verb by verb for global and grouped KPIs, and ggplot2 as a layered system (data, aesthetics, geometries, labels, themes). The week’s product: a KPI table and the 2–3 charts that communicate it.

Theory Guided practice Workshop Video · dplyr (part 1) Video · ggplot2 (part 2) Data: cafeteria.csv

Unit 2 — The analytics process and EDA

Week 5 — Proceso analítico y tipos de analítica. Business analytics as a process that turns data into actionable knowledge; translating business questions into analytical tasks; the workflow from question to decision; the roles in an analytics team. The in-class exercise feeds directly into Deliverable 1 of the final project.

Theory Podcast Guided practice Workshop

Week 6 — EDA: fuentes, limpieza y exploración. Data sources and quality; the diagnostic checklist — types, ranges, missing values, duplicates, inconsistent categories, outliers; cleaning with documented decisions; the reproducible raw → clean → analysis-ready pipeline.

Theory Podcast Guided practice Workshop Data: ferreteria_raw.csv

Week 7 — Examen Parcial 1. First integrative written exam (20%), covering weeks 1–6. September 7–12.

Unit 3 — AI applied to data analysis

Week 8 — Agentes de código: Claude Code, Cursor y VS Code. AI assistants for data analysis and programming; redoing with AI what was done “by hand” in the previous weeks (manipulation, cleaning, EDA, visualization); iterating prompts, reviewing generated code, verifying results.

Theory Guided practice Workshop Starter project (zip)

Week 9 — Presentación Avance del Proyecto (EDA). Deliverable 2 of the final project (5%): oral presentation of the exploratory analysis, with feedback from the professor and course assistants. September 21–26.

Unit 4 — Machine learning foundations

Week 10 — Fundamentos de Machine Learning. Machine learning as generalization; the standard pipeline — target and features, train/test split, metrics against a baseline; cross-validation; overfitting, data leakage, and badly chosen metrics. Two complementary videos build the confusion matrix and MAE/RMSE before the supervised weeks.

Theory Guided practice Workshop Video · Confusion matrix Video · RMSE & MAE Data: credito_taller10.csv · notas_taller10.csv · credito_evaluacion.csv · notas_evaluacion.csv

Unit 5 — Supervised learning

Week 11 — Clasificación: árboles y bosques. The full classification pipeline; the confusion matrix — why accuracy misleads, and the precision/recall trade-off; CART and random forests; choosing a model by the cost of its errors. The workshop predicts customer churn.

Theory Guided practice Workshop Video · Classification trees Data: credito_clasificacion.csv · clientes_conectatel.csv

Week 12 — Regresión: árboles, bosques y la ruta Lasso. The regression pipeline with MAE/RMSE against a baseline; regularized linear regression (Lasso), regression trees, and ensembles (random forest / XGBoost); interpretability vs. performance. The workshop prices apartments in Cali.

Theory Guided practice Workshop Video · Regression trees Data: notas_regresion.csv · apartamentos_cali.csv

Unit 6 — Unsupervised learning

Week 13 — Clustering: fundamentos y métricas. Clustering as segmentation without a target; k-means (distance, scaling, centroids); choosing k with the elbow and silhouette methods; profiling segments so they are actionable. The practice segments Spotify songs; the workshop segments gym members.

Theory Guided practice Workshop Video · k-means Data: spotify_canciones.csv · socios_califit.csv

Week 14 — Examen Parcial 2. Second integrative written exam (20%), covering weeks 8–13. October 26–31.

Week 15 — Simulacro de la presentación final. Each group presents a preliminary version of its final defense and receives feedback. November 2–7.

Week 16 — Presentación del Proyecto Final. Final oral defense of the project (Deliverable 3, 20%). It counts as an exam: attendance is mandatory. November 9–14.

Evaluation

Component	Weight	Allowed use of AI
Weekly concept quizzes (printed, at the start of class, every content week)	25%	No AI
In-class workshops (one per session, submitted on Intu)	10%	Collaboration with AI — every decision must be explainable
Integrative written exam 1 (week 7)	20%	No AI
Integrative written exam 2 (week 14)	20%	No AI
Final project (three deliverables)	25%	Planning / collaboration with AI, by deliverable

Quizzes and workshops have no make-ups; the worst quiz grade and the worst workshop grade are dropped once per semester. AI use in graded work follows ICESI’s institutional five-level IAG scale — each activity states its maximum allowed level, and every use must be declared, verified, and traceable.

Final project

Each group of 3 students receives a multi-variable business dataset — this semester, the Cóndor case, a fictional Andean payments-and-credit fintech —, formulates its own business question, and answers it with one of the course’s analytical tasks — classification, regression, or segmentation. Three deliverables: (1) the business question (week 6 — prerequisite for the later deliverables), (2) an oral progress presentation of the EDA (week 9, 5%), and (3) the final oral defense (week 16, 20%). There is no written document: the defense is the final product, and the presenting group member may be chosen at random.

Project guidelines (PDF, in Spanish) Cóndor case · context (PDF, in Spanish) Data (joined by cliente_id): base_clientes.csv · anexo_transacciones.csv · anexo_creditos.csv · anexo_campanas.csv

Bibliography and resources

• Wickham, H. & Grolemund, G. (2017). R for Data Science. O’Reilly Media.
• James, G., Witten, D., Hastie, T. & Tibshirani, R. (2021). An Introduction to Statistical Learning: with Applications in R (2nd ed.). Springer.
• Perkins, M., Roe, J. & Furze, L. (2025). Reimagining the Artificial Intelligence Assessment Scale (AIAS). JUTLP, 22(7).

---

Materials are released progressively as the semester advances. The course podcast — El Dato con Contexto — is on Spotify; quizzes are printed and answered at the start of each class, and workshops are submitted on Intu at the end of each session.