Categorización de Temperatura por Estado en México

Introducción

Este trabajo analiza la temperatura en los 32 estados de México entre 1985 y 2025, utilizando datos de temperatura mínima, máxima y media.

Mediante técnicas de clustering, los estados se agrupan en categorías climáticas (fríos, templados, cálidos) para identificar patrones regionales y comparar su diversidad térmica.

Instalación de Paqueterias y Llamado de Librerias

library(cluster)
library(ggplot2)
library(data.table)
library(factoextra)

## Welcome! Want to learn more? See two factoextra-related books at https://goo.gl/ve3WBa

library(dplyr)

## 
## Attaching package: 'dplyr'

## The following objects are masked from 'package:data.table':
## 
##     between, first, last

## The following objects are masked from 'package:stats':
## 
##     filter, lag

## The following objects are masked from 'package:base':
## 
##     intersect, setdiff, setequal, union

Llamado de base de datos

path <- "/Users/brisnaordaz/Downloads/data (1).csv" 
clima <- read.csv(path, stringsAsFactors = FALSE)

Datos entendidos

summary(clima)

##     PERIOD             STATE_ID     STATE               MIN_C      
##  Length:16071       Min.   : 0   Length:16071       Min.   :-2.20  
##  Class :character   1st Qu.: 8   Class :character   1st Qu.: 9.50  
##  Mode  :character   Median :16   Mode  :character   Median :14.10  
##                     Mean   :16                      Mean   :13.96  
##                     3rd Qu.:24                      3rd Qu.:18.60  
##                     Max.   :32                      Max.   :26.50  
##      MEAN_C          MAX_C       PRECIPITATION_MM     MIN_F      
##  Min.   : 7.50   Min.   :15.30   Min.   :  0.00   Min.   :28.04  
##  1st Qu.:17.60   1st Qu.:25.60   1st Qu.:  7.70   1st Qu.:49.10  
##  Median :21.70   Median :29.10   Median : 35.70   Median :57.38  
##  Mean   :21.39   Mean   :28.85   Mean   : 73.86   Mean   :57.12  
##  3rd Qu.:25.20   3rd Qu.:32.20   3rd Qu.:111.75   3rd Qu.:65.48  
##  Max.   :33.40   Max.   :40.40   Max.   :851.00   Max.   :79.70  
##      MEAN_F          MAX_F        PRECIPITATION_IN 
##  Min.   :45.50   Min.   : 59.54   Min.   : 0.0000  
##  1st Qu.:63.68   1st Qu.: 78.08   1st Qu.: 0.3031  
##  Median :71.06   Median : 84.38   Median : 1.4055  
##  Mean   :70.51   Mean   : 83.93   Mean   : 2.9080  
##  3rd Qu.:77.36   3rd Qu.: 89.96   3rd Qu.: 4.3996  
##  Max.   :92.12   Max.   :104.72   Max.   :33.5039

str(clima)

## 'data.frame':    16071 obs. of  11 variables:
##  $ PERIOD          : chr  "1985-01-01" "1985-01-01" "1985-01-01" "1985-01-01" ...
##  $ STATE_ID        : int  0 1 2 3 4 5 6 7 8 9 ...
##  $ STATE           : chr  "Mexico" "Aguascalientes" "Baja California" "Baja California Sur" ...
##  $ MIN_C           : num  7.8 3.1 5.6 9.2 15.7 2.9 14.4 14.1 -1.4 4.1 ...
##  $ MEAN_C          : num  15.9 12.2 12.9 17.1 22.7 10.3 22.7 20.9 7.5 12.5 ...
##  $ MAX_C           : num  23.9 21.3 20.2 25 29.7 17.7 30.9 27.8 16.5 20.9 ...
##  $ PRECIPITATION_MM: num  36 4.9 12.2 30.3 20.9 ...
##  $ MIN_F           : num  46 37.6 42.1 48.6 60.3 ...
##  $ MEAN_F          : num  60.6 54 55.2 62.8 72.9 ...
##  $ MAX_F           : num  75 70.3 68.4 77 85.5 ...
##  $ PRECIPITATION_IN: num  1.417 0.193 0.48 1.193 0.823 ...

Parseo de fecha y limpieza mínima

clima$PERIOD <- as.character(clima$PERIOD)

x <- as.Date(clima$PERIOD)                  # funciona si viene AAAA-MM-DD
bad <- is.na(x)
if (any(bad)) {
  x[bad] <- as.Date(paste0(clima$PERIOD[bad], "-01"))  # para AAAA-MM
}
clima$PERIOD <- x

clima <- subset(clima, STATE_ID != 0)
clima <- subset(clima, PERIOD >= as.Date("1985-01-01") & PERIOD <= as.Date("2025-07-01"))

Features por estado (temperatura)

features <- clima %>%
  group_by(STATE) %>%
  summarise(
    min_c_monthly_mean  = mean(MIN_C,  na.rm = TRUE),
    max_c_monthly_mean  = mean(MAX_C,  na.rm = TRUE),
    mean_c_monthly_mean = mean(MEAN_C, na.rm = TRUE)
  ) %>%
  mutate(thermal_amplitude = max_c_monthly_mean - min_c_monthly_mean) %>%
  ungroup()

head(features)

## # A tibble: 6 × 5
##   STATE               min_c_monthly_mean max_c_monthly_mean mean_c_monthly_mean
##   <chr>                            <dbl>              <dbl>               <dbl>
## 1 Aguascalientes                    8.91               26.4                17.6
## 2 Baja California                  12.0                26.8                19.4
## 3 Baja California Sur              15.5                30.3                22.9
## 4 Campeche                         20.3                33.1                26.7
## 5 Chiapas                          18.2                30.7                24.5
## 6 Chihuahua                         9.21               26.8                18.0
## # ℹ 1 more variable: thermal_amplitude <dbl>

Escalar los datos

X <- features %>% select(min_c_monthly_mean,
                         max_c_monthly_mean,
                         mean_c_monthly_mean,
                         thermal_amplitude)
X_scaled <- scale(X)

# Codo
fviz_nbclust(as.data.frame(X_scaled), kmeans, method = "wss") +
  ggtitle("Elección de K - Codo")

# Silueta
fviz_nbclust(as.data.frame(X_scaled), kmeans, method = "silhouette") +
  ggtitle("Elección de K - Silueta")

K-means

set.seed(42)
km4 <- kmeans(X_scaled, centers = 4, nstart = 50)
features$cluster_k4 <- factor(km4$cluster)

# Vista rápida del cluster plot
fviz_cluster(km4, data = as.data.frame(X_scaled), geom = "point",
             ellipse.type = "norm",
             main = "Clusters de temperatura por estado (K = 4)")

ord <- features %>%
  group_by(cluster_k4) %>%
  summarise(mean_temp = mean(mean_c_monthly_mean), .groups = "drop") %>%
  arrange(mean_temp) %>%
  mutate(label = c("Frío","Templado","Cálido","Muy cálido"))  # si cambias K, ajusta

features <- features %>%
  left_join(ord %>% select(cluster_k4, label), by = "cluster_k4")

Etiquetas interpretables

tabla_final <- features %>%
dplyr::arrange(mean_c_monthly_mean) %>%
dplyr::select(STATE,
min_c_monthly_mean,
max_c_monthly_mean,
mean_c_monthly_mean,
thermal_amplitude,
cluster_k4,
categoria = label)


tabla_final

## # A tibble: 32 × 7
##    STATE            min_c_monthly_mean max_c_monthly_mean mean_c_monthly_mean
##    <chr>                         <dbl>              <dbl>               <dbl>
##  1 Tlaxcala                       6.18               23.2                14.7
##  2 Estado de México               6.99               22.7                14.8
##  3 Ciudad de México               9.65               24.0                16.8
##  4 Zacatecas                      8.63               26.2                17.4
##  5 Hidalgo                        9.81               25.2                17.5
##  6 Aguascalientes                 8.91               26.4                17.6
##  7 Durango                        8.89               26.8                17.8
##  8 Puebla                        10.5                25.4                17.9
##  9 Chihuahua                      9.21               26.8                18.0
## 10 Guanajuato                    10.5                27.1                18.8
## # ℹ 22 more rows
## # ℹ 3 more variables: thermal_amplitude <dbl>, cluster_k4 <fct>,
## #   categoria <chr>

out_dir <- dirname(path)
write.csv(tabla_final, file.path(out_dir, "temperature_clusters_k4_1985_2025.csv"), row.names = FALSE)

Tabla final por estado

ggplot(tabla_final,
aes(x = reorder(STATE, mean_c_monthly_mean),
y = mean_c_monthly_mean, fill = categoria)) +
geom_col() +
coord_flip() +
labs(x = "Estado", y = "Temp. media (°C, 1985–2025)",
title = "Temperatura media por estado y categoría (K=4)") +
theme_minimal()