Visualization and Dashboard Gallery​‌​‌​‌​​‍​‌​​​‌​‌‍​​‌‌‌​‌​‍​​‌‌​‌​​‍​‌‌​​‌​‌‍​​‌‌‌​​​‍​‌‌​​‌​​‍​​‌‌‌​​‌‍​‌‌​​​‌​‍​​‌‌​​​‌‍​‌‌​​​​‌‍​​‌‌​‌​‌‍​‌‌​​‌‌​‍​​‌‌​‌‌​‍​‌‌​​‌​‌‍​​‌‌​‌‌‌‍​‌‌​​​‌‌‍​​‌‌​​‌‌‍​‌‌​​‌​​‍​​‌‌‌​‌​‍​​‌‌​​‌​‍​​‌‌​​​​‍​​‌‌​​‌​‍​​‌‌​‌‌​‍​​‌‌​​​​‍​​‌‌​​‌​‍​​‌‌​​​‌‍​​‌‌​​‌‌‍​​‌‌‌​‌​‍​‌‌​​‌​‌‍​‌‌​​‌​‌‍​​‌‌​‌‌‌‍​‌‌​​‌​‌‍​‌‌​​​​‌‍​‌‌​​‌‌​‍​​‌‌​‌​​‍​​‌‌​​​​¶

Results from the Machine Learning and NLP algorithms of the course.​‌​‌​‌​​‍​‌​​​‌​‌‍​​‌‌‌​‌​‍​​‌‌​‌​​‍​‌‌​​‌​‌‍​​‌‌‌​​​‍​‌‌​​‌​​‍​​‌‌‌​​‌‍​‌‌​​​‌​‍​​‌‌​​​‌‍​‌‌​​​​‌‍​​‌‌​‌​‌‍​‌‌​​‌‌​‍​​‌‌​‌‌​‍​‌‌​​‌​‌‍​​‌‌​‌‌‌‍​‌‌​​​‌‌‍​​‌‌​​‌‌‍​‌‌​​‌​​‍​​‌‌‌​‌​‍​​‌‌​​‌​‍​​‌‌​​​​‍​​‌‌​​‌​‍​​‌‌​‌‌​‍​​‌‌​​​​‍​​‌‌​​‌​‍​​‌‌​​​‌‍​​‌‌​​‌‌‍​​‌‌‌​‌​‍​‌‌​​‌​‌‍​‌‌​​‌​‌‍​​‌‌​‌‌‌‍​‌‌​​‌​‌‍​‌‌​​​​‌‍​‌‌​​‌‌​‍​​‌‌​‌​​‍​​‌‌​​​​ Each page includes analysis, charts, and source code.

Machine Learning¶

Full Analyses¶

• PCA + K-Means Clustering: Iris Dataset -- Dimensionality reduction and clustering on Fisher's classic dataset
• Manual PCA FactoMineR Style -- Replicating the gold standard of multivariate analysis in Python

Time Series¶

• ARIMA/SARIMA - Box-Jenkins Methodology -- Identification, estimation, diagnostics, and forecasting
• ARIMA Dashboard (interactive) -- 6 tabs: series, decomposition, ACF/PACF, diagnostics, forecast, radar
• ARIMA PRO Dashboard -- Bloomberg-style financial theme, KPIs, 7 tabs, model comparison
• ARIMA PRO Dashboard (interactive) -- Design inspired by OECD Explorer and Portfolio Optimizer

Computer Vision (Transfer Learning)¶

• Flower Classification with Transfer Learning -- MobileNetV2 + traditional ML on embeddings
• Flowers Dashboard (interactive) -- t-SNE, model comparison, confusion matrix, radar

Interactive Dashboards (HTML)¶

• PCA Iris Dashboard (interactive) -- Full-screen visualization
• FactoMineR Dashboard (interactive) -- Correlation circles and biplots

NLP and Text Mining¶

Guided Exercises¶

• Exercise 01: Text Anatomy and Word Counting -- Frequency, tokenization, and visualization
• Exercise 02: Stopword Filtering -- Semantic noise cleaning

Full Analyses¶

• Jaccard Similarity -- Document comparison and recommendation system
• Document Vectorization and Clustering -- TF-IDF, K-Means, and topic analysis

Panel Data Analysis (Big Data)¶

Full Pipeline: Spark + PostgreSQL + ML¶

• Module 06: QoG - 4 Research Lines + ML -- Central Asia, Water Security, Terrorism, Maghreb, PCA + K-Means
• QoG Dashboard (interactive) -- 5 tabs with interactive Plotly charts

Streaming and Cloud (Real-Time)¶

Live Dashboards¶

• Global Seismic Observatory -- Real-time earthquakes from USGS API
• Earthquake Dashboard (interactive) -- Leaflet map with live data, category filters
• ISS Tracker -- International Space Station tracking
• ISS Dashboard (interactive) -- Real-time position, pass predictor, orbital trajectory

Technologies¶

• Live APIs: USGS Earthquakes, Where The ISS At, Open Notify
• Visualization: Leaflet.js, CARTO dark tiles, Lucide icons
• Refresh rate: Auto-refresh every 5-30 seconds

​‌​‌​‌​​‍​‌​​​‌​‌‍​​‌‌‌​‌​‍​​‌‌​‌​​‍​‌‌​​‌​‌‍​​‌‌‌​​​‍​‌‌​​‌​​‍​​‌‌‌​​‌‍​‌‌​​​‌​‍​​‌‌​​​‌‍​‌‌​​​​‌‍​​‌‌​‌​‌‍​‌‌​​‌‌​‍​​‌‌​‌‌​‍​‌‌​​‌​‌‍​​‌‌​‌‌‌‍​‌‌​​​‌‌‍​​‌‌​​‌‌‍​‌‌​​‌​​‍​​‌‌‌​‌​‍​​‌‌​​‌​‍​​‌‌​​​​‍​​‌‌​​‌​‍​​‌‌​‌‌​‍​​‌‌​​​​‍​​‌‌​​‌​‍​​‌‌​​​‌‍​​‌‌​​‌‌‍​​‌‌‌​‌​‍​‌‌​​‌​‌‍​‌‌​​‌​‌‍​​‌‌​‌‌‌‍​‌‌​​‌​‌‍​‌‌​​​​‌‍​‌‌​​‌‌​‍​​‌‌​‌​​‍​​‌‌​​​​---

Source Code¶

The scripts that generate these visualizations are located in:

• ejercicios/04_machine_learning/ -- PCA, K-Means, Silhouette
• ejercicios/05_nlp_text_mining/ -- Counting, cleaning, sentiment, Jaccard
• ejercicios/06_análisis_datos_de_panel/ -- QoG Pipeline with Apache Spark
• ejercicios/07_infraestructura_bigdata/ -- Docker Compose, Spark Cluster
• ejercicios/08_streaming_kafka/ -- Kafka, Spark Structured Streaming
• ejercicios/09_cloud_localstack/ -- LocalStack, Terraform, AWS