Influence of the level of economic development on citizens eating habits
Mario Álvarez
2025-05-26
Introduction
This paper analyses the influence of the level of economic development on the dietary habits of citizens in various countries between 2010 and 2015. A dataset of economic, health and food consumption indicators is used to identify patterns and significant differences in diet according to the level of development.
The dataset covers information from 150 observations from 30 countries between 2010 and 2015, including economic variables such as GDP and the Human Development Index, health indicators and detailed measures of dietary consumption. This diversity of variables allows for a comprehensive analysis linking socio-economic context to dietary habits,
obj1 <- read.csv("cantcalesperanza_final2.csv")str(obj1)## 'data.frame': 150 obs. of 17 variables:
## $ Pais : chr "Austria" "Austria" "Austria" "Austria" ...
## $ Año : int 2010 2011 2012 2013 2014 2015 2010 2011 2012 2013 ...
## $ Estatus : chr "Developed" "Developed" "Developed" "Developed" ...
## $ Esperanza.de.vida: num 84 88 88 81.1 81.4 81.5 80 83 83 87 ...
## $ Mortalidad.adulta: int 75 73 7 68 66 65 81 8 78 77 ...
## $ Porcentaje.gastos: num 7423.2 8272.3 7878.4 90.6 8350.2 ...
## $ IDH : num 0.872 0.88 0.884 0.887 0.892 0.892 0.878 0.884 0.886 0.889 ...
## $ PIB : num 84 88 88 81.1 81.4 81.5 80 83 83 87 ...
## $ IMC : num 55.2 55.7 56.1 56.6 57.1 57.6 61.9 62.3 62.6 63 ...
## $ Total.kcal : int 3724 3735 3739 3768 3672 3664 3707 3720 3715 3733 ...
## $ Vegetal.kcal : int 2621 2632 2631 2631 2539 2503 2605 2572 2566 2561 ...
## $ Animal.kcal : int 1103 1104 1108 1137 1133 1161 1102 1147 1149 1172 ...
## $ GrasasAnim : num 263 259 281 269 233 ...
## $ Verduras : num 56.3 59.2 58.6 56.8 57.4 ...
## $ Fruits..other : num 40.5 43.3 40.2 43.1 38.8 ...
## $ Azucar : num 419 423 417 402 404 ...
## $ Cereales : num 15.1 23.2 27 18.9 18.3 ...summary(obj1)## Pais Año Estatus Esperanza.de.vida
## Length:150 Min. :2010 Length:150 Min. :72.40
## Class :character 1st Qu.:2011 Class :character 1st Qu.:76.33
## Mode :character Median :2012 Mode :character Median :81.10
## Mean :2012 Mean :80.23
## 3rd Qu.:2014 3rd Qu.:82.78
## Max. :2015 Max. :89.00
## Mortalidad.adulta Porcentaje.gastos IDH PIB
## Min. : 6.00 Min. : 0.00 Min. :0.7700 Min. :72.40
## 1st Qu.: 58.00 1st Qu.: 49.99 1st Qu.:0.8273 1st Qu.:76.33
## Median : 76.00 Median : 758.91 Median :0.8600 Median :81.10
## Mean : 82.35 Mean : 2196.88 Mean :0.8586 Mean :80.23
## 3rd Qu.:116.25 3rd Qu.: 2655.76 3rd Qu.:0.8888 3rd Qu.:82.78
## Max. :188.00 Max. :17028.53 Max. :0.9260 Max. :89.00
## IMC Total.kcal Vegetal.kcal Animal.kcal
## Min. : 6.00 Min. :2640 Min. :1907 Min. : 632.0
## 1st Qu.:57.62 1st Qu.:3166 1st Qu.:2172 1st Qu.: 825.2
## Median :59.70 Median :3368 Median :2362 Median : 938.0
## Mean :53.07 Mean :3311 Mean :2350 Mean : 960.3
## 3rd Qu.:62.58 3rd Qu.:3486 3rd Qu.:2515 3rd Qu.:1090.8
## Max. :69.60 Max. :3804 Max. :2677 Max. :1427.0
## GrasasAnim Verduras Fruits..other Azucar
## Min. : 2.87 Min. : 26.72 Min. :11.91 Min. :141.5
## 1st Qu.: 43.91 1st Qu.: 49.82 1st Qu.:26.61 1st Qu.:239.7
## Median : 96.94 Median : 58.20 Median :31.11 Median :295.9
## Mean :108.02 Mean : 60.88 Mean :33.09 Mean :309.4
## 3rd Qu.:139.24 3rd Qu.: 67.89 3rd Qu.:40.69 3rd Qu.:377.5
## Max. :288.65 Max. :109.73 Max. :61.61 Max. :543.9
## Cereales
## Min. : 0.000
## 1st Qu.: 2.708
## Median : 7.240
## Mean :10.275
## 3rd Qu.:12.318
## Max. :76.510To facilitate the analysis, derived variables were created that synthesise the energy composition of the diet, such as the percentage of calories from animal and plant sources, the ratio between the two and an index of dietary diversity.
# Comparo PIB e IMC por estatus de desarrollo
obj1 <- obj1 %>%
mutate(
Pct_Animal = 100 * Animal.kcal / Total.kcal,
Pct_Vegetal = 100 * Vegetal.kcal / Total.kcal,
Ratio_A_V = Animal.kcal / Vegetal.kcal,
Umbral = 0.05 * Total.kcal,
Diversidad_Alim = (Vegetal.kcal > Umbral) + (Animal.kcal > Umbral) + (GrasasAnim > Umbral) +
(Verduras > Umbral) + (`Fruits..other` > Umbral) + (Azucar > Umbral) + (Cereales > Umbral)
) %>%
select(-Umbral)
head(obj1)## Pais Año Estatus Esperanza.de.vida Mortalidad.adulta Porcentaje.gastos
## 1 Austria 2010 Developed 84.0 75 7423.22877
## 2 Austria 2011 Developed 88.0 73 8272.30676
## 3 Austria 2012 Developed 88.0 7 7878.37236
## 4 Austria 2013 Developed 81.1 68 90.58501
## 5 Austria 2014 Developed 81.4 66 8350.19352
## 6 Austria 2015 Developed 81.5 65 0.00000
## IDH PIB IMC Total.kcal Vegetal.kcal Animal.kcal GrasasAnim Verduras
## 1 0.872 84.0 55.2 3724 2621 1103 263.23 56.33
## 2 0.880 88.0 55.7 3735 2632 1104 258.55 59.17
## 3 0.884 88.0 56.1 3739 2631 1108 281.08 58.56
## 4 0.887 81.1 56.6 3768 2631 1137 268.56 56.81
## 5 0.892 81.4 57.1 3672 2539 1133 232.54 57.40
## 6 0.892 81.5 57.6 3664 2503 1161 242.55 50.91
## Fruits..other Azucar Cereales Pct_Animal Pct_Vegetal Ratio_A_V
## 1 40.49 418.97 15.09 29.61869 70.38131 0.4208317
## 2 43.29 422.66 23.18 29.55823 70.46854 0.4194529
## 3 40.20 416.84 27.02 29.63359 70.36641 0.4211326
## 4 43.12 401.71 18.95 30.17516 69.82484 0.4321551
## 5 38.82 404.49 18.34 30.85512 69.14488 0.4462387
## 6 41.45 424.03 11.49 31.68668 68.31332 0.4638434
## Diversidad_Alim
## 1 4
## 2 4
## 3 4
## 4 4
## 5 4
## 6 4summary(obj1)## Pais Año Estatus Esperanza.de.vida
## Length:150 Min. :2010 Length:150 Min. :72.40
## Class :character 1st Qu.:2011 Class :character 1st Qu.:76.33
## Mode :character Median :2012 Mode :character Median :81.10
## Mean :2012 Mean :80.23
## 3rd Qu.:2014 3rd Qu.:82.78
## Max. :2015 Max. :89.00
## Mortalidad.adulta Porcentaje.gastos IDH PIB
## Min. : 6.00 Min. : 0.00 Min. :0.7700 Min. :72.40
## 1st Qu.: 58.00 1st Qu.: 49.99 1st Qu.:0.8273 1st Qu.:76.33
## Median : 76.00 Median : 758.91 Median :0.8600 Median :81.10
## Mean : 82.35 Mean : 2196.88 Mean :0.8586 Mean :80.23
## 3rd Qu.:116.25 3rd Qu.: 2655.76 3rd Qu.:0.8888 3rd Qu.:82.78
## Max. :188.00 Max. :17028.53 Max. :0.9260 Max. :89.00
## IMC Total.kcal Vegetal.kcal Animal.kcal
## Min. : 6.00 Min. :2640 Min. :1907 Min. : 632.0
## 1st Qu.:57.62 1st Qu.:3166 1st Qu.:2172 1st Qu.: 825.2
## Median :59.70 Median :3368 Median :2362 Median : 938.0
## Mean :53.07 Mean :3311 Mean :2350 Mean : 960.3
## 3rd Qu.:62.58 3rd Qu.:3486 3rd Qu.:2515 3rd Qu.:1090.8
## Max. :69.60 Max. :3804 Max. :2677 Max. :1427.0
## GrasasAnim Verduras Fruits..other Azucar
## Min. : 2.87 Min. : 26.72 Min. :11.91 Min. :141.5
## 1st Qu.: 43.91 1st Qu.: 49.82 1st Qu.:26.61 1st Qu.:239.7
## Median : 96.94 Median : 58.20 Median :31.11 Median :295.9
## Mean :108.02 Mean : 60.88 Mean :33.09 Mean :309.4
## 3rd Qu.:139.24 3rd Qu.: 67.89 3rd Qu.:40.69 3rd Qu.:377.5
## Max. :288.65 Max. :109.73 Max. :61.61 Max. :543.9
## Cereales Pct_Animal Pct_Vegetal Ratio_A_V
## Min. : 0.000 Min. :22.68 Min. :57.22 Min. :0.2935
## 1st Qu.: 2.708 1st Qu.:25.99 1st Qu.:69.00 1st Qu.:0.3512
## Median : 7.240 Median :28.48 Median :71.50 Median :0.3983
## Mean :10.275 Mean :28.92 Mean :71.08 Mean :0.4116
## 3rd Qu.:12.318 3rd Qu.:31.00 3rd Qu.:74.01 3rd Qu.:0.4494
## Max. :76.510 Max. :42.81 Max. :77.29 Max. :0.7483
## Diversidad_Alim
## Min. :2.000
## 1st Qu.:3.000
## Median :3.000
## Mean :3.133
## 3rd Qu.:3.000
## Max. :4.000resumen <- obj1 %>%
group_by(Estatus) %>%
summarise(
Media_Pct_Animal = mean(Pct_Animal, na.rm = TRUE),
Mediana_Pct_Animal = median(Pct_Animal, na.rm = TRUE),
SD_Pct_Animal = sd(Pct_Animal, na.rm = TRUE),
Media_Pct_Vegetal = mean(Pct_Vegetal, na.rm = TRUE),
Mediana_Pct_Vegetal = median(Pct_Vegetal, na.rm = TRUE),
SD_Pct_Vegetal = sd(Pct_Vegetal, na.rm = TRUE),
Media_Ratio_A_V = mean(Ratio_A_V, na.rm = TRUE),
Mediana_Ratio_A_V = median(Ratio_A_V, na.rm = TRUE),
SD_Ratio_A_V = sd(Ratio_A_V, na.rm = TRUE),
Media_Diversidad = mean(Diversidad_Alim, na.rm = TRUE),
Mediana_Diversidad = median(Diversidad_Alim, na.rm = TRUE),
SD_Diversidad = sd(Diversidad_Alim, na.rm = TRUE)
)
print(resumen)## # A tibble: 2 × 13
## Estatus Media_Pct_Animal Mediana_Pct_Animal SD_Pct_Animal Media_Pct_Vegetal
## <chr> <dbl> <dbl> <dbl> <dbl>
## 1 Developed 28.5 28.2 3.24 71.5
## 2 Developing 30.9 31.7 6.47 69.1
## # ℹ 8 more variables: Mediana_Pct_Vegetal <dbl>, SD_Pct_Vegetal <dbl>,
## # Media_Ratio_A_V <dbl>, Mediana_Ratio_A_V <dbl>, SD_Ratio_A_V <dbl>,
## # Media_Diversidad <dbl>, Mediana_Diversidad <dbl>, SD_Diversidad <dbl>summary(resumen)## Estatus Media_Pct_Animal Mediana_Pct_Animal SD_Pct_Animal
## Length:2 Min. :28.54 Min. :28.23 Min. :3.240
## Class :character 1st Qu.:29.14 1st Qu.:29.10 1st Qu.:4.047
## Mode :character Median :29.74 Median :29.97 Median :4.854
## Mean :29.74 Mean :29.97 Mean :4.854
## 3rd Qu.:30.34 3rd Qu.:30.84 3rd Qu.:5.661
## Max. :30.94 Max. :31.71 Max. :6.468
## Media_Pct_Vegetal Mediana_Pct_Vegetal SD_Pct_Vegetal Media_Ratio_A_V
## Min. :69.07 Min. :68.29 Min. :3.241 Min. :0.4023
## 1st Qu.:69.66 1st Qu.:69.16 1st Qu.:4.047 1st Qu.:0.4169
## Median :70.26 Median :70.03 Median :4.853 Median :0.4315
## Mean :70.26 Mean :70.03 Mean :4.853 Mean :0.4315
## 3rd Qu.:70.86 3rd Qu.:70.90 3rd Qu.:5.660 3rd Qu.:0.4460
## Max. :71.46 Max. :71.77 Max. :6.466 Max. :0.4606
## Mediana_Ratio_A_V SD_Ratio_A_V Media_Diversidad Mediana_Diversidad
## Min. :0.3933 Min. :0.06532 Min. :3.000 Min. :3
## 1st Qu.:0.4111 1st Qu.:0.08448 1st Qu.:3.040 1st Qu.:3
## Median :0.4288 Median :0.10364 Median :3.079 Median :3
## Mean :0.4288 Mean :0.10364 Mean :3.079 Mean :3
## 3rd Qu.:0.4466 3rd Qu.:0.12281 3rd Qu.:3.119 3rd Qu.:3
## Max. :0.4643 Max. :0.14197 Max. :3.159 Max. :3
## SD_Diversidad
## Min. :0.0000
## 1st Qu.:0.1201
## Median :0.2401
## Mean :0.2401
## 3rd Qu.:0.3602
## Max. :0.4802To analyse the differences in dietary habits according to the level of development, comparisons of the derived variables were made between countries categorised as developed and developing. The boxplots clearly showed that developing countries tend to have a higher percentage of calories from animal sources and sugar, while developed countries have a higher percentage of vegetable calories and animal fats. This disparity reflects distinctive dietary patterns linked to economic and cultural factors. In addition, the bar chart breaking down average calorie intake by food type and economic status highlighted these differences quantitatively, facilitating the identification of clear trends that could guide specific nutritional policies for each group.
# Gráfico de barras: consumo calórico por tipo de alimento y estatus de desarrollo
df_long <- obj1 %>%
select(Estatus, Vegetal.kcal, Animal.kcal, Azucar, GrasasAnim, Verduras) %>%
group_by(Estatus) %>%
summarise(across(where(is.numeric), mean, na.rm = TRUE)) %>%
pivot_longer(cols = -Estatus, names_to = "Grupo", values_to = "kcal")## Warning: There was 1 warning in `summarise()`.
## ℹ In argument: `across(where(is.numeric), mean, na.rm = TRUE)`.
## ℹ In group 1: `Estatus = "Developed"`.
## Caused by warning:
## ! The `...` argument of `across()` is deprecated as of dplyr 1.1.0.
## Supply arguments directly to `.fns` through an anonymous function instead.
##
## # Previously
## across(a:b, mean, na.rm = TRUE)
##
## # Now
## across(a:b, \(x) mean(x, na.rm = TRUE))ggplot(df_long, aes(x = Grupo, y = kcal, fill = Estatus)) +
geom_col(position = "dodge") +
labs(title = "Consumo promedio por grupo alimenticio según nivel de desarrollo", x = "Grupo alimenticio", y = "kcal promedio") +
theme_minimal() +
theme(axis.text.x = element_text(angle = 45, hjust = 1))
# Boxplot % kcal animal
p1 <-ggplot(obj1, aes(x = Estatus, y = Pct_Animal, fill = Estatus)) +
geom_boxplot() +
labs(
x = "Nivel de desarrollo económico",
y = "% kcal de origen animal") +
theme_minimal() +
theme(legend.position = "none")
# Boxplot % kcal vegetal
p2 <-ggplot(obj1, aes(x = Estatus, y = Pct_Vegetal, fill = Estatus)) +
geom_boxplot() +
labs(
x = "Nivel de desarrollo económico",
y = "% kcal de origen vegetal") +
theme_minimal() +
theme(legend.position = "none")
# Boxplot Ratio Animal / Vegetal
p3 <-ggplot(obj1, aes(x = Estatus, y = Ratio_A_V, fill = Estatus)) +
geom_boxplot() +
labs(
x = "Nivel de desarrollo económico",
y = "Ratio kcal animal / vegetal") +
theme_minimal() +
theme(legend.position = "none")
grid.arrange(p1, p2, p3, nrow = 1)
Analysis of the temporal evolution of diet provided insights into how dietary habits have changed between 2010 and 2015. It was observed that the proportion of animal calories has shown a moderate increase in both groups, with a more pronounced increase in developing countries. This finding suggests that the nutrition transition, with increased consumption of animal products, is more active in developing countries, possibly due to economic improvements or changes in cultural preferences. These temporal trends are essential to anticipate future public health challenges and to adjust appropriate intervention strategies.
obj1 %>%
group_by(Año) %>%
summarise(esperanza_media = mean(Esperanza.de.vida, na.rm = TRUE)) %>%
ggplot(aes(x = Año, y = esperanza_media)) +
geom_line(color = "blue") +
labs(title = "Evolución de la Esperanza de Vida Promedio", y = "Esperanza de Vida", x = "Año") +
theme_minimal()
wilcox.test(Pct_Animal ~ Estatus, data = obj1)##
## Wilcoxon rank sum test with continuity correction
##
## data: Pct_Animal by Estatus
## W = 1297, p-value = 0.2715
## alternative hypothesis: true location shift is not equal to 0To quantify the influence of economic and health variables on dietary habits, linear regression models were fitted. In these models, the percentage of animal calories and dietary diversity were considered as dependent variables, while the Human Development Index, Gross Domestic Product per capita, body mass index and year were included as predictors. The results showed that HDI and GDP are significant variables that explain important variations in diet, confirming their relevance as determinants of dietary behaviour at the population level. In addition, the inclusion of BMI allowed controlling for possible influences related to the health status of the population, enriching the interpretation of the models.
g1 <-ggscatter(obj1, x = "IDH", y = "Pct_Animal",
add = "reg.line", conf.int = TRUE,
cor.coef = TRUE, cor.method = "pearson",
xlab = "IDH", ylab = "% kcal animal",
title = "Relación entre IDH y % kcal animal")
g2 <-ggscatter(obj1, x = "IDH", y = "Pct_Vegetal",
add = "reg.line", conf.int = TRUE,
cor.coef = TRUE, cor.method = "pearson",
xlab = "IDH", ylab = "% kcal vegetal",
title = "Relación entre IDH y % kcal vegetal")
obj1 <- obj1 %>% mutate(log_PIB = log(PIB + 1)) # Añadimos 1 para evitar log(0)
g3 <-ggscatter(obj1, x = "log_PIB", y = "Pct_Animal",
add = "reg.line", conf.int = TRUE,
cor.coef = TRUE, cor.method = "pearson",
xlab = "Log(PIB)", ylab = "% kcal animal",
title = "Relación entre log(PIB) y % kcal animal")
grid.arrange(g1,g2,g3, nrow = 1)
modelo1 <- lm(Pct_Animal ~ IDH + Año, data = obj1)
summary(modelo1)##
## Call:
## lm(formula = Pct_Animal ~ IDH + Año, data = obj1)
##
## Residuals:
## Min 1Q Median 3Q Max
## -6.2082 -2.5066 -0.2937 2.2880 12.0830
##
## Coefficients:
## Estimate Std. Error t value Pr(>|t|)
## (Intercept) 122.19201 329.12276 0.371 0.711
## IDH 57.75426 7.42021 7.783 1.16e-12 ***
## Año -0.07099 0.16400 -0.433 0.666
## ---
## Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
##
## Residual standard error: 3.388 on 147 degrees of freedom
## Multiple R-squared: 0.294, Adjusted R-squared: 0.2844
## F-statistic: 30.61 on 2 and 147 DF, p-value: 7.722e-12modelo2 <- lm(Diversidad_Alim ~ log_PIB + IMC + Año, data = obj1)
summary(modelo2)##
## Call:
## lm(formula = Diversidad_Alim ~ log_PIB + IMC + Año, data = obj1)
##
## Residuals:
## Min 1Q Median 3Q Max
## -1.13261 -0.19418 -0.14910 0.08001 0.92684
##
## Coefficients:
## Estimate Std. Error t value Pr(>|t|)
## (Intercept) 30.128168 41.875079 0.719 0.47300
## log_PIB -0.881454 0.662975 -1.330 0.18574
## IMC 0.005214 0.001818 2.867 0.00475 **
## Año -0.011626 0.020913 -0.556 0.57913
## ---
## Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
##
## Residual standard error: 0.4333 on 146 degrees of freedom
## Multiple R-squared: 0.06565, Adjusted R-squared: 0.04645
## F-statistic: 3.419 on 3 and 146 DF, p-value: 0.01902Principal Component Analysis was used to reduce the complexity of the set of derived food variables, allowing the identification of the underlying factors that best explain the variability in dietary habits across countries. Through this technique, it was possible to visualise how countries are distributed in a reduced space of components, where the first components accumulate most of the relevant information. The clear differentiation between developed and developing countries observed in the biplot graphs confirms that the selected variables capture substantial differences in dietary patterns, facilitating a simpler and more robust interpretation of the existing relationships.
alim_vars <- obj1 %>%
select(Pct_Animal, Pct_Vegetal, Ratio_A_V, Diversidad_Alim) %>%
na.omit()
pca <- PCA(alim_vars, scale.unit = TRUE, graph = FALSE)
fviz_pca_biplot(pca, repel = TRUE)
obj1_clean <- obj1 %>%
filter(!is.na(Pct_Animal), !is.na(Pct_Vegetal), !is.na(Ratio_A_V), !is.na(Diversidad_Alim), !is.na(Estatus))
estatus_pca <- factor(obj1_clean$Estatus)
fviz_pca_ind(pca,
geom.ind = "point",
habillage = estatus_pca,
addEllipses = TRUE,
palette = "jco",
legend.title = "Nivel de desarrollo")
obj1 %>%
group_by(Año, Estatus) %>%
summarise(Media_Pct_Animal = mean(Pct_Animal, na.rm = TRUE), .groups = "drop") %>%
ggplot(aes(x = Año, y = Media_Pct_Animal, color = Estatus)) +
geom_line(linewidth = 1) +
labs(title = "% kcal animal promedio anual por nivel de desarrollo",
y = "% kcal animal", x = "Año") +
theme_minimal()
It was used to group countries according to the similarity of their dietary profiles, based on previously standardised derived variables. Ward’s method, recognised for its effectiveness in minimising the internal variance of the groups, allowed the construction of a dendrogram reflecting the hierarchical structure of the clusters. Natural clusters within the sample were identified as showing high internal cohesion, and comparison with the classification according to the level of economic development revealed a strong correspondence. This technique provided a complementary and visual insight that reinforces the hypothesis that economic development influences eating habits in a consistent and structured way.
# Uso las variables derivadas normalizadas
alim_norm <- scale(obj1_clean %>%
select(Pct_Animal, Pct_Vegetal, Ratio_A_V, Diversidad_Alim))
# Calculo distancia y agrupación jerárquica
distancia <- dist(alim_norm)
p1 <- fviz_nbclust(alim_norm, hcut, method = "silhouette",
hc_method = "ward.D2", k.max = 10) + labs(title = "Número óptimo de clusters (Silhouette)")
p2 <- fviz_nbclust(alim_norm, hcut, method = "wss",
hc_method = "ward.D2", k.max = 10) + labs(title = "Número óptimo de clusters (Codo)")
grid.arrange(p1, p2, nrow = 1)
# Según criterio, selecciono número de clusters 3
num_clusters <- 3
# Clustering jerárquico con método Ward
clust <- hclust(distancia, method = "ward.D2")
# Corto dendrograma en num_clusters grupos
grupos <- cutree(clust, k = num_clusters)
# Añado clusters a datos originales
obj1_clean$Cluster <- factor(grupos)
# Visualizo dendrograma con clusters coloreados y rectángulos
fviz_dend(clust, k = num_clusters,
cex = 0.6, # tamaño texto etiquetas
color_labels_by_k = TRUE,
rect = TRUE, rect_border = "jco", rect_fill = TRUE,
show_labels = FALSE,
main = "Dendrograma de patrones alimentarios")## Warning: The `<scale>` argument of `guides()` cannot be `FALSE`. Use "none" instead as
## of ggplot2 3.3.4.
## ℹ The deprecated feature was likely used in the factoextra package.
## Please report the issue at <https://github.com/kassambara/factoextra/issues>.
## This warning is displayed once every 8 hours.
## Call `lifecycle::last_lifecycle_warnings()` to see where this warning was
## generated.
# Comparo clusters con Estatus
table(obj1_clean$Cluster, obj1_clean$Estatus)##
## Developed Developing
## 1 24 0
## 2 84 12
## 3 18 12It provided an essential tool for examining bivariate relationships between economic and diet-derived variables. Using scatter plots with fitted regression lines and correlation matrices, statistically significant associations were identified that corroborate and complement previous findings. This graphical and numerical analysis provided clarity on the strength and direction of the relationships, helping to validate the hypotheses put forward and reinforce the final conclusions of the study.
anova_animal <- aov(Pct_Animal ~ Estatus, data = obj1_clean)
summary(anova_animal)## Df Sum Sq Mean Sq F value Pr(>F)
## Estatus 1 115.8 115.80 7.535 0.0068 **
## Residuals 148 2274.4 15.37
## ---
## Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1anova_diversidad <- aov(Diversidad_Alim ~ Estatus, data = obj1_clean)
summary(anova_diversidad)## Df Sum Sq Mean Sq F value Pr(>F)
## Estatus 1 0.508 0.5079 2.608 0.108
## Residuals 148 28.825 0.1948modelo_multi <- lm(Pct_Animal ~ IDH + PIB + IMC + Año, data = obj1_clean)
summary(modelo_multi)##
## Call:
## lm(formula = Pct_Animal ~ IDH + PIB + IMC + Año, data = obj1_clean)
##
## Residuals:
## Min 1Q Median 3Q Max
## -6.062 -2.529 -0.091 2.056 11.885
##
## Coefficients:
## Estimate Std. Error t value Pr(>|t|)
## (Intercept) 71.79256 322.08123 0.223 0.8239
## IDH 73.88916 9.94247 7.432 8.53e-12 ***
## PIB -0.20806 0.08563 -2.430 0.0163 *
## IMC -0.02702 0.01386 -1.949 0.0532 .
## Año -0.04382 0.16053 -0.273 0.7853
## ---
## Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
##
## Residual standard error: 3.302 on 145 degrees of freedom
## Multiple R-squared: 0.3387, Adjusted R-squared: 0.3204
## F-statistic: 18.56 on 4 and 145 DF, p-value: 2.444e-12pairs(obj1_clean %>%
select(Pct_Animal, Pct_Vegetal, Ratio_A_V, Diversidad_Alim, IDH, PIB, IMC),
main = "Matriz de correlaciones (visual)")
##Conclusions
This study has clearly and rigorously demonstrated how the level of economic development of countries profoundly influences the dietary habits of their citizens. Through a comprehensive analysis combining socio-economic variables, health and detailed dietary measures, it has shown that differences in dietary composition and diversity are closely linked to the economic context.
Derived variables capturing the percentage of calories from animal and plant sources, as well as an index of dietary diversity, revealed differential dietary patterns: developing countries tend to consume a higher proportion of calories from animal foods and sugars, while developed countries have diets with a higher intake of plant calories and animal fats, reflecting deep cultural and economic influences.
The combined use of multivariate techniques such as Principal Component Analysis and hierarchical clustering allowed not only to reduce the complexity of the data but also to identify natural groupings of countries with homogeneous dietary profiles, strongly aligned with their level of development. This finding provides strong and visual support for the central hypothesis of the study.
The fitted regression models confirmed the crucial role of the Human Development Index and Gross Domestic Product per capita as significant predictors of dietary patterns, evidencing the interaction between economic well-being and dietary choices. In addition, the incorporation of body mass index as a control variable enriched the understanding of the phenomenon from a multidimensional perspective.
Overall, this work provides relevant empirical evidence for policy makers, nutritionists and economists, highlighting the need to address dietary habits from a comprehensive approach that considers the socio-economic context. The clear differentiation and dynamics detected between countries suggest that strategies to improve nutrition should be adapted to specific economic conditions, with special attention to the nutrition transition in developing countries.
Finally, this research opens the door to future lines of study, including deeper longitudinal analyses, incorporation of cultural and regional factors, and exploration of interventions that promote healthier and more sustainable diets as a function of economic development. Undoubtedly, this report represents a significant contribution to the field of socio-economic analysis of food and lays the foundation for future research that will contribute to improving global health through informed and contextualised policies.