Skip to content
Merged
Show file tree
Hide file tree
Changes from all commits
Commits
File filter

Filter by extension

Filter by extension

Conversations
Failed to load comments.
Loading
Jump to
Jump to file
Failed to load files.
Loading
Diff view
Diff view
1 change: 1 addition & 0 deletions .github/smoke/smoke.js
Original file line number Diff line number Diff line change
Expand Up @@ -54,6 +54,7 @@ const PAGES = [
{ id: 'ia-compute', path: '/ia-compute/' },
{ id: 'tasas', path: '/tasas-cambio/' },
{ id: 'poblacion', path: '/poblacion-latam/' },
{ id: 'adopcion-ia', path: '/adopcion-ia/' },
{ id: 'redir-pulso', path: '/pulso-financiero/', redirect: '/claude/pulso-financiero/' },
{ id: 'redir-sla', path: '/sla-operaciones/', redirect: '/claude/sla-operaciones/' },
].filter(p => !ONLY || ONLY.includes(p.id));
Expand Down
2 changes: 2 additions & 0 deletions README.en.md
Original file line number Diff line number Diff line change
Expand Up @@ -20,6 +20,7 @@ The whole site is **bilingual ES/EN** with a single switcher in the top bar. The
/ia-compute/ → Explainer · The acceleration of AI compute (1958–2026)
/tasas-cambio/ → Explainer · Official central-bank exchange rates (Mexico, Colombia, Argentina, Brazil, Chile, Uruguay, Peru)
/poblacion-latam/ → Explainer · Official population of Latin America (20 countries, series and projections)
/adopcion-ia/ → Template · Enterprise AI adoption: four questions, eight metrics (illustrative case with hypothetical data)
```
The **ChatGPT Work** portfolio lives on an external site: `https://javier-dashboards.jforero.chatgpt.site/` (its English version at `/en`).

Expand Down Expand Up @@ -59,6 +60,7 @@ Real data from cited sources. Every figure states its level of evidence.
| **The acceleration of AI compute · 1958–2026** | 68 years of training compute, from Rosenblatt's artificial neuron to the frontier models of 2026. 72 models, each with its **stated level of evidence** (published · confirmed · likely · speculative · trend). Logarithmic scale, the EU AI Act systemic-risk threshold (10²⁶ FLOP) and an explicit contrast with Moore's Law. | [Epoch AI](https://epoch.ai/data/ai-models) · CC BY | [`/ia-compute/`](./ia-compute/) |
| **Official exchange rates · Mexico, Colombia, Argentina, Brazil, Chile, Uruguay and Peru** | How much is each currency worth against the others today, and how far is it from its recent behaviour? A list with flags and the latest official value of each series, next to the history of the chosen rate, full daily history with weekends and holidays flagged, a base-100 comparison, changes and a converter. **No server:** it reads the data repository's JSON files live (raw.githubusercontent with a jsDelivr fallback), and every figure carries its date and source. Cross rates between MXN, COP, ARS, BRL, CLP, UYU and PEN are labelled **calculated**. **Resources (in Spanish):** [YouTube video «Hice este Dashboard con Claude Design (Sin Ser Diseñador)»](https://youtu.be/zOTZKeb0lCs) · [article «Claude Dashboards no reemplaza tus tableros: pone a prueba tus datos»](https://javierforero.co/2026/10/10/claude-dashboards-analitica-conversacional/). | [Bank of Mexico](https://www.banxico.org.mx/SieAPIRest/service/v1/) (FIX) · [Banco de la República](https://suameca.banrep.gov.co/estadisticas-economicas/) (TRM) · [BCRA](https://www.bcra.gob.ar/archivos/Catalogo/Content/files/pdf/estadisticascambiarias-v1.pdf) (A 3500) · [BCB](https://olinda.bcb.gov.br/olinda/servico/PTAX/versao/v1/documentacao) (PTAX) · [BCCh](https://si3.bcentral.cl/estadisticas/Principal1/Web_Services/documentacion.html) (observed dollar) · [BCU](https://www.bcu.gub.uy/Estadisticas-e-Indicadores/Paginas/Cotizaciones.aspx) (interbank rate) · [BCRP](https://estadisticas.bcrp.gob.pe/estadisticas/series/ayuda/api) (interbank selling rate) · processed in [`fx-bancos-centrales`](https://github.com/jaforero/fx-bancos-centrales) | [`/tasas-cambio/`](./tasas-cambio/) |
| **Official population of Latin America · series and projections** | How many people live in the region, according to those mandated to count them? In tabs: **the regional total of the 20 countries as a reference**, a map with a year selector, a country series with visible breaks, **population pyramids of each country against the region**, the aging index, a base-100 index, growth and peak-population year. **No server:** it reads the data repository's JSON files live (raw.githubusercontent with jsDelivr backup). Every figure carries its institution, vintage and raw file; the UN serves only as a declared fallback, and countries without a verifiable source stay **pending, never filled in**. | National institutes · [UN WPP 2024](https://population.un.org/wpp/) 2024 Rev. as fallback · processed in [`latam-population-official`](https://github.com/jaforero/latam-population-official) | [`/poblacion-latam/`](./poblacion-latam/) |
| **AI adoption · four questions, eight metrics** | Does your company have AI, use it, build it into its processes and get results? Eight metrics computed from system records, never from surveys: access, activation, habitual use, process coverage, workflow integration, outcome against a comparison group, quality and net time. Every card opens its formula, its rows and an aggregated download. **It is also a template:** five CSV files the reader fills in and loads; they are processed in the browser and never sent or stored. No composite index, no traffic lights, no per-person data. **Resource:** [article «Medir la adopción de IA: no es un porcentaje, son cuatro preguntas»](https://javierforero.co/2026/11/14/medir-adopcion-ia-kpis/) (Spanish and English on the same page). | **Illustrative case with hypothetical data** (Northgate Supply, a distributor that does not exist) · author's own framework, not a standard | [`/adopcion-ia/`](./adopcion-ia/) |

## Forecasting methodology (Financial Pulse)

Expand Down
2 changes: 2 additions & 0 deletions README.md
Original file line number Diff line number Diff line change
Expand Up @@ -20,6 +20,7 @@ Todo el sitio es **bilingüe ES/EN** con un selector único en la barra superior
/ia-compute/ → Divulgación · La aceleración del cómputo en IA (1958–2026)
/tasas-cambio/ → Divulgación · Tasas de cambio oficiales de bancos centrales (México, Colombia, Argentina, Brasil, Chile, Uruguay, Perú)
/poblacion-latam/ → Divulgación · Población oficial de América Latina (20 países, series y proyecciones)
/adopcion-ia/ → Plantilla · Adopción de IA en la empresa: cuatro preguntas, ocho indicadores (caso ilustrativo con datos hipotéticos)
```
El portafolio de **ChatGPT Work** vive en un sitio externo: `https://javier-dashboards.jforero.chatgpt.site/` (su versión inglesa, en `/en`).

Expand Down Expand Up @@ -59,6 +60,7 @@ Datos reales de fuentes citadas. Cada cifra declara su nivel de evidencia.
| **La aceleración del cómputo en IA · 1958–2026** | 68 años de cómputo de entrenamiento, de la neurona artificial de Rosenblatt a los modelos frontera de 2026. 72 modelos, cada uno con su **nivel de evidencia declarado** (publicado · confirmado · probable · especulativo · tendencia). Escala logarítmica, umbral de riesgo sistémico de la EU AI Act (10²⁶ FLOP) y contraste explícito con la Ley de Moore. | [Epoch AI](https://epoch.ai/data/ai-models) · CC BY | [`/ia-compute/`](./ia-compute/) |
| **Tasas de cambio oficiales · México, Colombia, Argentina, Brasil, Chile, Uruguay y Perú** | ¿Cuánto vale hoy cada moneda frente a las otras y qué tan lejos está de su comportamiento reciente? Lista con banderas y el último valor oficial de cada serie, junto al histórico de la tasa elegida, histórico diario completo con fines de semana y festivos diferenciados, comparación base 100, variaciones y conversor. **Sin servidor:** lee en vivo los JSON del repositorio de datos (raw.githubusercontent con respaldo jsDelivr), y cada cifra lleva su fecha y su fuente. Las cruzadas entre MXN, COP, ARS, BRL, CLP, UYU y PEN se rotulan como **calculadas**. **Recursos:** [video en YouTube «Hice este Dashboard con Claude Design (Sin Ser Diseñador)»](https://youtu.be/zOTZKeb0lCs) · [artículo «Claude Dashboards no reemplaza tus tableros: pone a prueba tus datos»](https://javierforero.co/2026/10/10/claude-dashboards-analitica-conversacional/). | [Banco de México](https://www.banxico.org.mx/SieAPIRest/service/v1/) (FIX) · [Banco de la República](https://suameca.banrep.gov.co/estadisticas-economicas/) (TRM) · [BCRA](https://www.bcra.gob.ar/archivos/Catalogo/Content/files/pdf/estadisticascambiarias-v1.pdf) (A 3500) · [BCB](https://olinda.bcb.gov.br/olinda/servico/PTAX/versao/v1/documentacao) (PTAX) · [BCCh](https://si3.bcentral.cl/estadisticas/Principal1/Web_Services/documentacion.html) (dólar observado) · [BCU](https://www.bcu.gub.uy/Estadisticas-e-Indicadores/Paginas/Cotizaciones.aspx) (interbancario) · [BCRP](https://estadisticas.bcrp.gob.pe/estadisticas/series/ayuda/api) (interbancario venta) · procesados en [`fx-bancos-centrales`](https://github.com/jaforero/fx-bancos-centrales) | [`/tasas-cambio/`](./tasas-cambio/) |
| **Población oficial de América Latina · series y proyecciones** | ¿Cuánta gente vive en la región, según quién tiene el mandato de contarla? En pestañas: **total regional de los 20 países como referencia**, mapa con selector de año, serie por país con rupturas visibles, **pirámides de población de cada país contra la región**, índice de envejecimiento, índice base 100, crecimiento y año de máxima población. **Sin servidor:** lee en vivo los JSON del repositorio de datos (raw.githubusercontent con respaldo jsDelivr). Cada cifra lleva institución, vintage y archivo crudo; la ONU solo como respaldo declarado y los países sin fuente verificable quedan **pendientes, nunca rellenados**. | Institutos nacionales · [ONU WPP 2024](https://population.un.org/wpp/) Rev. 2024 de respaldo · procesados en [`latam-population-official`](https://github.com/jaforero/latam-population-official) | [`/poblacion-latam/`](./poblacion-latam/) |
| **Adopción de IA · cuatro preguntas, ocho indicadores** | ¿Tu empresa tiene IA, la usa, la integró en sus procesos y obtiene resultados? Ocho indicadores calculados desde registros de sistemas, nunca desde encuestas: acceso, activación, uso habitual, cobertura de procesos, integración en el flujo, resultado con grupo de comparación, calidad y tiempo neto. Cada tarjeta abre su fórmula, sus filas y una descarga agregada. **Es también una plantilla:** cinco CSV que el lector llena y carga; se procesan en su navegador y no se envían ni se guardan. Sin índice compuesto, sin semáforos y sin datos por persona. **Recurso:** [artículo «Medir la adopción de IA: no es un porcentaje, son cuatro preguntas»](https://javierforero.co/2026/11/14/medir-adopcion-ia-kpis/). | **Caso ilustrativo con datos hipotéticos** (Aurora, una distribuidora que no existe) · marco de elaboración propia, no es un estándar | [`/adopcion-ia/`](./adopcion-ia/) |

## Metodología predictiva (Pulso Financiero)

Expand Down
61 changes: 61 additions & 0 deletions adopcion-ia/README.en.md
Original file line number Diff line number Diff line change
@@ -0,0 +1,61 @@
[Español](README.md) · **English**

# AI adoption scorecard · four questions, eight metrics

[![Live piece](https://img.shields.io/badge/live-piece-4e00ff)](https://dashboards.javierforero.co/adopcion-ia/?lang=en)
[![MIT License](https://img.shields.io/badge/license-MIT-041c59)](../LICENSE)
[![No dependencies](https://img.shields.io/badge/JavaScript-no%20dependencies-0048ff)](#files)
[![Bilingual](https://img.shields.io/badge/ES%20%C2%B7%20EN-bilingual-7c4dff)](https://dashboards.javierforero.co/adopcion-ia/)

The question it answers: **does your company have AI, use it, build it into its processes and get results?** Those are four different facts, and none follows from the one before. It goes with the article ["Measuring AI adoption: it is not one percentage, it is four questions"](https://javierforero.co/2026/11/14/medir-adopcion-ia-kpis/).

> **Illustrative case with hypothetical data.** Northgate Supply (Aurora in Spanish) is a distributor that does not exist. Its numbers are the ones in the article. Framework and scorecard are the author's own work: this is not a standard and it has not been comparatively validated.

![AI adoption scorecard](assets/og-image.png)

## Two uses
1. **Browsable example.** The case from the article, with the same numbers.
2. **Working template.** Download five CSV templates, fill them with your company's records, load them and get your own metrics. **Your files are processed in your browser. They are not sent to any server and they are not stored.**

## Design principle: traceable facts, not perceptions
- No metric uses surveys or self-reported data. Vendor "assisted hours" and "estimated value" are ignored if a file includes them.
- Every card has **"See where it comes from"**: formula, numerator, denominator, window, file, row count and a download of the aggregated subset (never per person).
- **No data, no number.** If a file or a column is missing, the card says "No data" and which record is missing. Nothing is imputed or filled with zeros.
- No composite index, no traffic lights, no outside targets, no per-person data. No value for groups of fewer than 5 people.

## Metrics
| # | Metric | Formula | File |
|---|---|---|---|
| 1 | Access | eligible with access ÷ eligible | people |
| 2 | Activation | with access and ≥ 1 action in 28 days ÷ with access | people, weekly_usage |
| 3 | Habitual use | active in ≥ 9 of 12 weeks ÷ with ≥ 1 active week (also over eligible) | people, weekly_usage |
| 4 | Process coverage | with a verified use case (≥ 2 pieces of evidence) ÷ where AI is applicable | processes |
| 5 | Workflow integration | Σ assisted ÷ Σ closed, last 4 weeks of the "after" period for the group with AI | units |
| 6 | Operating outcome against a comparison | (after − before with AI) − (after − before in comparison) | units |
| 7 | Quality and rework | the same double subtraction on the rework rate, in percentage points | units |
| 8 | Net time recovered | (drop in drafting minutes − rise in review minutes) × assisted units in the last 4 weeks | units |

Alerts (not metrics): **concentration** (actions from the most active 10%) and **use outside the tools you provide** (`alerts.csv`).

Templates accept Spanish or English column names (`data/plantillas/` and `data/plantillas/en/`).

## Files
| File | Role |
|---|---|
| `index.html` | Page and styles (site tokens, light and dark theme) |
| `calc.js` | Dependency-free engine: CSV, validation, the eight metrics, alerts and readings. Runs the same in the browser and in Node |
| `app.js` | Screens, hand-built SVG charts, file loading and downloads |
| `i18n.js` | ES/EN strings by key |
| `data/ejemplo/` | The five CSV files of the illustrative case and `etiquetas-en.js` (same data, English labels) |
| `data/plantillas/` | Templates in Spanish, `en/` in English, and column dictionaries |
| `tools/generar-datos.js` | Generates all of `data/` from a fixed seed |
| `tools/probar.js` | Recomputes the eight values from the CSV files and fails if any does not match |

## Tests
```bash
node adopcion-ia/tools/generar-datos.js # regenerates data/ (deterministic)
node adopcion-ia/tools/probar.js # 54 checks; exit code 1 if any fails
```

## Privacy and analytics
GA4 records the visit, as on the rest of the site. Switching to "My data" sets `ga-disable` before any file is read. Loaded data lives in memory only: no `fetch`, `localStorage`, cookies or IndexedDB touch it. Only the `jf-lang` and `jf-theme` preferences are stored.
Loading
Loading