Building the heatmap

The script below (training.R) reads the 2025 sheet from running.xlsx, cleans up the dates and distances, and then lays each day out on a small monthly calendar grid faceted by month. Let’s walk through it piece by piece.

1. Load the libraries and read the log. Nothing fancy here — readxl pulls the spreadsheet in, dplyr and lubridate do the date wrangling, and ggplot2 draws the thing.

training.R

library(ggplot2)
library(dplyr)
library(lubridate)
library(scales)
library(readxl)  # For reading Excel files

# Read the Excel file
kms <- read_excel("blog/2026/09/running.xlsx", sheet = "2025")

2. Clean the data and print a quick summary. Dates get coerced properly, rows with missing values are dropped, and distances are forced to numeric. Nothing here should ever trigger, but a training log accumulated by hand over a year will eventually have a typo’d cell.

training.R

kms <- kms %>%
  mutate(date = as.Date(date)) %>%
  filter(!is.na(date), !is.na(kms)) %>%
  mutate(kms = as.numeric(kms)) %>%
  filter(kms >= 0)

cat("Data Summary:\n")
cat("Date range:", as.character(min(kms$date)), "to", as.character(max(kms$date)), "\n")
cat("Total days:", nrow(kms), "\n")
cat("Total kilometers:", sum(kms$kms, na.rm = TRUE), "km\n")
cat("Average daily kilometers:", round(mean(kms$kms, na.rm = TRUE), 2), "km\n")

3. Work out where each day sits on its monthly calendar grid. ggplot2 has no built-in calendar geometry, so this is the part that actually earns its keep: for every date we compute its weekday (Monday-first) and its week_of_month, correcting for the fact that the 1st of the month rarely falls on a Monday. Each day also gets tagged with that month’s running total, which ends up in the facet label.

training.R

create_kms_calendar_heatmap <- function(data, title = "Daily Kilometers Calendar Heatmap") {

  calendar_data_prep <- data %>%
    rename(value = kms)

  monthly_totals <- calendar_data_prep %>%
    mutate(year = year(date), month = month(date)) %>%
    group_by(year, month) %>%
    summarise(total_kms = sum(value, na.rm = TRUE), .groups = 'drop')

  calendar_data <- calendar_data_prep %>%
    mutate(
      year = year(date),
      month = month(date),
      day = day(date),
      weekday = wday(date, week_start = 1), # Monday = 1
      month_name = month(date, label = TRUE, abbr = TRUE)
    ) %>%
    group_by(year, month) %>%
    mutate(
      first_day_of_month = floor_date(date, "month"),
      first_weekday = wday(first_day_of_month, week_start = 1),
      # Offset by how far into the week the 1st of the month falls
      week_of_month = ceiling((day + first_weekday - 1) / 7)
    ) %>%
    ungroup() %>%
    left_join(monthly_totals, by = c("year", "month")) %>%
    mutate(
      month_year_label = paste0(month_name, " ", year, "\n(", round(total_kms, 1), " km)"),
      month_year_date = as.Date(paste(year, month, "01", sep = "-"))
    ) %>%
    arrange(month_year_date)

  month_order <- calendar_data %>%
    select(month_year_label, month_year_date) %>%
    distinct() %>%
    arrange(month_year_date) %>%
    pull(month_year_label)

  calendar_data$month_year_label <- factor(calendar_data$month_year_label, levels = month_order)

  calendar_data
}

4. Draw the tiles. With the calendar grid in place, the plot itself is a straightforward geom_tile() faceted by month, colored on a diverging scale centered on the median distance — so an easy 5K and a 25K long run stand out on opposite ends, and typical training days fade into the gray middle.

training.R

plot_kms_calendar <- function(calendar_data, title = "Daily Kilometers Calendar Heatmap") {
  ggplot(calendar_data, aes(x = weekday, y = -week_of_month, fill = value)) +
    geom_tile(color = "white", linewidth = 0.5) +
    facet_wrap(~ month_year_label, ncol = 4, scales = "free_y") +
    scale_x_continuous(
      breaks = 1:7,
      labels = c("Mon", "Tue", "Wed", "Thu", "Fri", "Sat", "Sun"),
      position = "top"
    ) +
    scale_y_continuous(breaks = NULL) +
    scale_fill_gradient2(
      low = "lightgray", mid = "yellow", high = "darkred",
      midpoint = median(calendar_data$value, na.rm = TRUE),
      name = "Kilometers"
    ) +
    labs(title = title, x = "", y = "") +
    theme_minimal() +
    theme(
      axis.text.y = element_blank(),
      axis.text.x.bottom = element_blank(),
      axis.ticks = element_blank(),
      panel.grid = element_blank(),
      strip.background = element_blank(),
      legend.position = "bottom"
    )
}

kms_calendar_plot <- plot_kms_calendar(create_kms_calendar_heatmap(kms))
print(kms_calendar_plot)

5. A colorblind-friendly variant, and the monthly numbers behind the chart. Swapping in the viridis “plasma” scale is a one-liner. And since the heatmap is meant to replace staring at a spreadsheet, it’s only fair to also print the spreadsheet-style summary it’s replacing.

training.R

kms_calendar_viridis <- plot_kms_calendar(create_kms_calendar_heatmap(kms), "Daily Kilometers") +
  scale_fill_viridis_c(name = "Kilometers", option = "plasma", trans = "sqrt")
print(kms_calendar_viridis)

monthly_stats <- kms %>%
  mutate(year_month = floor_date(date, "month")) %>%
  group_by(year_month) %>%
  summarise(
    total_kms = sum(kms, na.rm = TRUE),
    avg_daily_kms = mean(kms, na.rm = TRUE),
    max_daily_kms = max(kms, na.rm = TRUE),
    days_with_data = n(),
    .groups = 'drop'
  )

print(monthly_stats)

And here’s the result for 2025:

The gaps are the honest part of the chart — the weeks I meant to run and didn’t are just as visible as the weeks I did. All told, 2025 added up to 5,310 km — enough to have jogged from Córdoba to Ulaanbaatar, if I’d only picked a less roundabout route than “out my front door and back.”

What is Regression Analysis?

Regression analysis is a statistical method used to predict the value of a dependent variable based on the values of one or more independent variables. It involves analyzing the relationship between the dependent variable and the independent variables to understand how changes in the independent variables affect the dependent variable.

The independent variables, also known as predictors, are used to estimate or “predict” the value of the dependent variable. The relationship between the variables is typically represented by an equation or a mathematical model. The regression model estimates the relationship by fitting a line (or curve) through the data points, allowing for the prediction of the dependent variable for given values of the independent variables.

Regression analysis can be used to measure the influence of one or multiple variables on another variable. By identifying the strength and direction of the relationship, regression analysis helps researchers and analysts to understand the underlying factors that impact the dependent variable. It is a powerful tool for making predictions, forecasting future outcomes, and understanding the complex mechanisms driving the variables of interest.

Regression analysis finds applications in various fields such as economics, finance, social sciences, and healthcare. In economics, it can be used to analyze the relationship between factors like income, inflation, and consumer spending. In finance, it aids in predicting stock prices and identifying factors that impact investment returns. In the social sciences, regression analysis is used to study the impact of variables like education, income, and demographic characteristics on various outcomes. In healthcare, it helps in understanding the factors influencing patient outcomes and optimizing treatment strategies.

Regression analysis can also assist in model validation and hypothesis testing. Researchers can develop different regression models to test alternative hypotheses and evaluate the significance of each variable. The quality of the regression model can be assessed by examining various statistical measures such as and p-values. These measures provide information about the goodness-of-fit of the model and the statistical significance of the coefficients.

In the field of statistics, regression analysis is a widely used method for modeling the relationship between a dependent variable and one or more independent variables. It helps in understanding the nature and strength of the relationship between variables, making predictions, and drawing inferences. There are various types of regression analysis, each suitable for different scenarios and data types. In this post, we will explore the simple and multiple linear regression, however, another common types of regression models are: logistic regression, ordinal regression, and nominal regression among others.

The motive behind linear regression

Linear regression is useful when we suspect a linear relationship between variables (known as explanatory variables, predictors, or covariates) and a response variable. While it may seem straightforward, linear regression is a powerful and widely used statistical learning method.

Matplotlib is building the font cache; this may take a moment.

(-10.0, 310.0)

(0.0, 28.27)

Testing the significance of a regression

There are several ways the significance of a regression can be tested. Providing errors are normally and identically distributed, a parametric test can be used. Analysis of Variance (ANOVA) is often the preferred approach, although one can also use a t-test to test whether the slope is significantly different from zero. If errors are not normally and identically distributed, then a randomization test should be used.

1. Introduction to the XGBoost algorithm

XGBoost, (Extreme Gradient Boosting), is one of the most widely used supervised machine learning algorithms that uses the boosting principle.

Supervised learning is one that has input variables and an output variable and an output variable and uses an algorithm to learn the mapping function of the input to the output.The objective is to approximate the mapping function as accurately as possible so that when you have new input data , the output variables can be predicted for that data.

It is called supervised learning because the process of an algorithm learning from the training data set can be thought of as a teacher supervising the learning process. We know the correct answers, the algorithm iteratively makes predictions on the training data and is corrected by the teacher. Learning stops when the algorithm reaches an acceptable level of performance.

Supervised learning problems can be grouped into regression and classification problems.

Classification: A classification problem is when the output variable is a category, such as “red” or “blue” or “disease” and “no disease”.

Regression: A regression problem is when the output variable is a real value, such as “euros” or “kilograms”. Some common types of problems built on classification and regression include recommendation and time series prediction, respectively. Some popular examples of supervised machine learning algorithms are:

Linear regression for regression problems.

Random forest for classification and regression problems.

Support vector machine for classification problems.

The xgboost algorithm is similar to gradient boosting but more efficient. It has linear model solutions and tree learning algorithms and is at least 10 times faster than existing implementations of gradient boosting. It supports several objective functions, including regression, classification and ranking. and what makes it fast is its ability to do parallel computations on a single machine.

In terms of efficiency, accuracy and feasibility it is more powerful than the random forest algorithm, for example or a neural network and since it has very high predictive power but relatively slow implementation, xgboost becomes a suitable choice for solving most regression, classification and ranking problems as well as user-created objective functions. It also has additional features to do cross validation and find the most important variables, which makes it interesting for credit scoring models, insurance claims, or where there are many parameters to be controlled to optimize the model.

As can be seen in the following graph (1), the XGBoost model has the best combination of prediction performance and processing time compared to other algorithms.

Chollet and Allaire, (2018) (2) summarize the value of XGBoost as follows:

“XGBoost is used for problems where there is the availability of structured data is very wide, while deep learning is used for perceptual problems such as image classification. Users of the former almost always use the XGBoost library.”

“These are the two techniques that one should be most familiar with to be successful in applied machine learning today: gradient boosting machines, for shallow learning problems; and deep learning, for perceptual problems. In technical terms, this means that one should be familiar with XGBoost and Keras.”

Originally, XGBoost is a library written in C++ and exported to R in the xgboost package in our case, the XGBoost model has been trained using the package in R caret(3).

2. Objectives and data preparation

In this post and preliminary preview of the project with data application to the Spanish real estate market we train and tune a model based on the XGBoost algorithm using the tidymodels library of R. For this purpose and as data source we use the AmesHousing dataset (4) which contains data of 82 variables for 2,930 properties in Ames County, Iowa. Our model will attempt to predict the selling price of the home.

# Data load
library(AmesHousing)

# Cleaning and data preparation libraries
library(janitor)
library(dplyr)

# Libraries needed
library(rsample)
library(recipes)
library(parsnip)
library(tune)
library(dials)
library(workflows)
library(yardstick)

# acceleration of calculations with parallel processing (optional but useful)
library(doParallel)
all_cores <- parallel::detectCores(logical = FALSE)
registerDoParallel(cores = all_cores)

# random seed set.seed() for results to be replicable
# 
set.seed(1234)

# data load and cleaning names
ames_data <- make_ames() %>%
  janitor::clean_names()

3. XGBoost Process

# Data split for training and testing. Stratification by sales price
ames_split <- rsample::initial_split(
  ames_data, 
  prop = 0.8, 
  strata = sale_price
)

# preprocessing routine
preprocessing_recipe <- 
  recipes::recipe(sale_price ~ ., data = training(ames_split)) %>%
  # categorical variables to factor variables conversion
  recipes::step_string2factor(all_nominal()) %>%
  # low frequency factor levels combination
  recipes::step_other(all_nominal(), threshold = 0.01) %>%
  # remove of predictors without variance that do not provide predicitive information
  recipes::step_nzv(all_nominal()) %>%
  prep()

ames_cv_folds <- 
  recipes::bake(
    preprocessing_recipe, 
    new_data = training(ames_split)
  ) %>%  
  rsample::vfold_cv(v = 5)

# Model specification
xgboost_model <- 
  parsnip::boost_tree(
    mode = "regression",
    trees = 1000,
    min_n = tune(),
    tree_depth = tune(),
    learn_rate = tune(),
    loss_reduction = tune()
  ) %>%
    set_engine("xgboost", objective = "reg:squarederror")

# Grid setup
xgboost_params <- 
  dials::parameters(
    min_n(),
    tree_depth(),
    learn_rate(),
    loss_reduction()
  )

xgboost_grid <- 
  dials::grid_max_entropy(
    xgboost_params, 
    size = 60
  )

knitr::kable(head(xgboost_grid))

xgboost_wf <- 
  workflows::workflow() %>%
  add_model(xgboost_model) %>% 
  add_formula(sale_price ~ .)

# hyperparameters tuning
xgboost_tuned <- tune::tune_grid(
  object = xgboost_wf,
  resamples = ames_cv_folds,
  grid = xgboost_grid,
  metrics = yardstick::metric_set(rmse, rsq, mae),
  control = tune::control_grid(verbose = TRUE)
)

xgboost_tuned %>%
  tune::show_best(metric = "rmse") %>%
  knitr::kable()

xgboost_best_params <- xgboost_tuned %>%
  tune::select_best("rmse")

knitr::kable(xgboost_best_params)

xgboost_model_final <- xgboost_model %>% 
  finalize_model(xgboost_best_params)

train_processed <- bake(preprocessing_recipe,  new_data = training(ames_split))

train_prediction <- xgboost_model_final %>%
  # model fit on all training data
  fit(
    formula = sale_price ~ ., 
    data    = train_processed
  ) %>%
  # prediction of sales prices on training data
  predict(new_data = train_processed) %>%
  bind_cols(training(ames_split))

xgboost_score_train <- 
  train_prediction %>%
  yardstick::metrics(sale_price, .pred) %>%
  mutate(.estimate = format(round(.estimate, 2), big.mark = ","))

knitr::kable(xgboost_score_train)

test_processed  <- bake(preprocessing_recipe, new_data = testing(ames_split))

test_prediction <- xgboost_model_final %>%
  # model fit on all training data
  fit(
    formula = sale_price ~ ., 
    data    = train_processed
  ) %>%
  # use of the training fit model for test data prediction
  predict(new_data = test_processed) %>%
  bind_cols(testing(ames_split))

# measuring the accuracy of our model using `yardstick`.
xgboost_score <- 
  test_prediction %>%
  yardstick::metrics(sale_price, .pred) %>%
  mutate(.estimate = format(round(.estimate, 2), big.mark = ","))

knitr::kable(xgboost_score)

house_prediction_residual <- test_prediction %>%
  arrange(.pred) %>%
  mutate(residual_pct = (sale_price - .pred) / .pred) %>%
  select(.pred, residual_pct)

ggplot(house_prediction_residual, aes(x = .pred, y = residual_pct)) +
  geom_point() +
  xlab("Predicted Sale Price") +
  ylab("Residual (%)") +
  scale_x_continuous(labels = scales::dollar_format()) +
  scale_y_continuous(labels = scales::percent)

Conclusions

The objective of this analysis was to work through the training process of an XGBoost model in R using the tidymodels package, and to learn the basics of how the algorithm works, although we have not put too much emphasis on the performance of our model, the foundations for future lines of research with this tool have been laid.

We have seen that the tidymodels library provides us with a standard process and tools to handle resampling (rsample), data preprocessing (recipes), model specification (parsnip), fitting (tune) and model validation (yardstick). In this sense, the tidymodels ability to sort the machine learning process is a step-change improvement for machine learning accessibility in R; thus, it is easier to train and understand the XGBoost model training process.

References

1. https://towardsdatascience.com/https-medium-com-vishalmorde-xgboost-algorithm-long-she-may-rein-edd9f99be63d

2. Chollet, F. and Allaire, J. J. (2018): “Deep Learning with R,” Ed. Manning Publications.

3. Kuhn, M. (2008). “Building Predictive Models in R Using the caret Package,” Journal of Statistical Software, 28(5), 1-26; doi:http://dx.doi.org/10.18637/jss.v028.i05.

4. De Cock, D. (2011). “Ames, Iowa: Alternative to the Boston Housing Data as an End of Semester Regression Project,”: Journal of Statistics Education, Volume 19, Number 3.

5. Chen, T. and Guestrin, C. (2016): “XGBoost: A Scalable Tree Boosting System,” doi:10.1145/2939672.2939785

6. Friedman, J.H. (2001): “Greedy Function Approximation: A Gradient Boosting Machine,” Annals of Statistics, pp. 1189–1232.

Introduction

Bibliographic Data

The bibliographic data are based on the search query that has been made. For the moment we have focused on one of the lines of the SEJ670 UCO Research Group, neophilia/neophobia in tourism. Needles to say that it would be applied to other research lines; however, watch out when filtering and redefining the database because we have to remove areas that are not of our interest, for example, in our selection there were references about zoology, biology, etc…, given that there is also neophilia and neophobia, especially the latter, specifically in animals, well, we must be careful because some area can sneak in the form of journals and references not related to our research area.

Data source: Clarivate Analytics Web of Science (http://apps.webofknowledge.com)

Format: Bibtex

Query: “Web of Science Core Collection”

Range: 1995-2022

Document type: All

Query date: September, 2022

Package installation, data loading and conversion

As the objective is to see what analysis the tool does and how it does it, the steps of installation, data loading and conversion have been omitted, although it is an important step since it must be done in R, after downloading the data from WoS or Scopus, which also has its own technique. After a lot of research we managed to find a way to merge both databases and it works perfectly, and it also automatically omits duplicate references. With this procedure we can have both databases in a single file and although WoS has more references (as a general rule, although not always), Scopus almost always has some different one/s and even if they are few, at least we practically cover the two most important sources of articles. It is worth mentioning that in the last updates of this Bibliometrix package other open bibliographic databases such as OpenAlex, Dimensions, Lens, and from the medical field such as PubMed and Cochrane have been included.

1. Descriptive Analysis

The descriptive analysis of the package provides a lot of information on the annual development of research, the most productive authors, articles, countries and relevant keywords.

MAIN INFORMATION ABOUT DATA

 Timespan                              1995 : 2022 
 Sources (Journals, Books, etc)        69 
 Documents                             152 
 Annual Growth Rate %                  9.64 
 Document Average Age                  9.52 
 Average citations per doc             29.24 
 Average citations per year per doc    2.425 
 References                            7867 
 
DOCUMENT TYPES                     
 article                         130 
 article; book chapter           2 
 article; early access           5 
 article; proceedings paper      1 
 editorial material              4 
 letter                          1 
 proceedings paper               3 
 review                          6 
 
DOCUMENT CONTENTS
 Keywords Plus (ID)                    617 
 Author's Keywords (DE)                512 
 
AUTHORS
 Authors                               467 
 Author Appearances                    531 
 Authors of single-authored docs       18 
 
AUTHORS COLLABORATION
 Single-authored docs                  19 
 Documents per Author                  0.325 
 Co-Authors per Doc                    3.49 
 International co-authorships %        36.18 
 

Annual Scientific Production

 Year    Articles
    1995        1
    2000        2
    2003        2
    2005        3
    2006        2
    2007        3
    2008        5
    2009        3
    2010        7
    2011        2
    2012        6
    2013        5
    2014        8
    2015       11
    2016       11
    2017       13
    2018       13
    2019       11
    2020       17
    2021       10
    2022       12

Annual Percentage Growth Rate 9.64 


Most Productive Authors

     Authors        Articles   Authors        Articles Fractionalized
1  METTKE-HOFMANN C        8 METTKE-HOFMANN C                    3.53
2  BUGNYAR T               4 STHAPIT E                           1.33
3  EVES A                  4 EVES A                              1.20
4  MORAND-FERRON J         4 MORAND-FERRON J                     1.20
5  BISAZZA A               3 [ANONYMOUS] A                       1.00
6  CARACCIOLO F            3 AKYUZ BG                            1.00
7  GRIFFIN AS              3 ANTONAKIS J                         1.00
8  KIM YG                  3 ARMELAGOS GJ                        1.00
9  VERNEAU F               3 BERTI I                             1.00
10 WIDDIG A                3 CHAKRABARTI B                       1.00


Top manuscripts per citations

                                 Paper                                      DOI  TC TCperYear  NTC
1  KIM YG, 2009, INT J HOSP MANAG               10.1016/j.ijhm.2008.11.005      346     20.35 2.38
2  CHANG RCY, 2010, ANN TOURIS RES              10.1016/j.annals.2010.03.007    276     17.25 3.21
3  ANTONAKIS J, 2017, LEADERSH Q                10.1016/j.leaqua.2017.01.006    192     21.33 4.79
4  VERBEKE W, 2005, BR FOOD J                   10.1108/00070700510629779       183      8.71 1.97
5  KIM YG, 2010, INT J HOSP MANAG               10.1016/j.ijhm.2009.10.015      178     11.12 2.07
6  CAVIGELLI SA, 2003, PROC NATL ACAD SCI U S A 10.1073/pnas.2535721100         161      7.00 1.22
7  HUGHES RN, 2007, NEUROSCI BIOBEHAV REV       10.1016/j.neubiorev.2006.11.004 110      5.79 2.84
8  MORAND-FERRON J, 2011, BEHAV ECOL            10.1093/beheco/arr120           106      7.07 1.49
9  DAY RL, 2003, ANIM BEHAV                     10.1006/anbe.2003.2074          103      4.48 0.78
10 REVERDY C, 2008, APPETITE                    10.1016/j.appet.2008.01.010     101      5.61 1.51


Corresponding Author's Countries

          Country Articles   Freq SCP MCP MCP_Ratio
1  UNITED KINGDOM       24 0.1622  17   7     0.292
2  USA                  18 0.1216  12   6     0.333
3  CHINA                14 0.0946   5   9     0.643
4  AUSTRALIA            13 0.0878   9   4     0.308
5  GERMANY               9 0.0608   4   5     0.556
6  CANADA                8 0.0541   5   3     0.375
7  ITALY                 7 0.0473   7   0     0.000
8  AUSTRIA               6 0.0405   2   4     0.667
9  BELGIUM               3 0.0203   1   2     0.667
10 BRAZIL                3 0.0203   2   1     0.333


SCP: Single Country Publications

MCP: Multiple Country Publications


Total Citations per Country

      Country      Total Citations Average Article Citations
1  UNITED KINGDOM             1122                     46.75
2  USA                         639                     35.50
3  CHINA                       490                     35.00
4  AUSTRALIA                   275                     21.15
5  BELGIUM                     193                     64.33
6  SWITZERLAND                 192                     96.00
7  AUSTRIA                     187                     31.17
8  GERMANY                     158                     17.56
9  ARGENTINA                   143                     71.50
10 ITALY                       143                     20.43


Most Relevant Sources

                                    Sources        Articles
1  ANIMAL BEHAVIOUR                                      17
2  ETHOLOGY                                               8
3  ANIMAL COGNITION                                       7
4  FOOD QUALITY AND PREFERENCE                            7
5  BEHAVIORAL ECOLOGY                                     6
6  BRITISH FOOD JOURNAL                                   6
7  SCIENTIFIC REPORTS                                     6
8  APPETITE                                               5
9  INTERNATIONAL JOURNAL OF HOSPITALITY MANAGEMENT        5
10 PLOS ONE                                               5


Most Relevant Keywords

   Author Keywords (DE)      Articles Keywords-Plus (ID)     Articles
1             NEOPHILIA            24           NEOPHOBIA          48
2             FOOD NEOPHOBIA       23           BEHAVIOR           29
3             NEOPHOBIA            22           NEOPHILIA          21
4             EXPLORATION          13           ATTITUDES          17
5             FOOD                 11           SCALE              17
6             FOOD TOURISM         11           SATISFACTION       16
7             FOOD NEOPHILIA       10           TOURISM            15
8             PERSONALITY           9           EVOLUTION          13
9             INNOVATION            8           INFORMATION        13
10            LOCAL FOOD            8           MODEL              13

                                                                                   [,1]
PLINER P, 1992, APPETITE, V19, P105, DOI 10.1016/0195-6663(92)90014-W                55
GREENBERG R, 2001, CURR ORNITHOL, V16, P119                                          32
COHEN E, 2004, ANN TOURISM RES, V31, P755, DOI 10.1016/J.ANNALS.2004.02.003          29
RITCHEY PN, 2003, APPETITE, V40, P163, DOI 10.1016/S0195-6663(02)00134-4             27
METTKE-HOFMANN C, 2002, ETHOLOGY, V108, P249, DOI 10.1046/J.1439-0310.2002.00773.X   24
TUORILA H, 2001, FOOD QUAL PREFER, V12, P29, DOI 10.1016/S0950-3293(00)00025-2       24
KIM YG, 2009, INT J HOSP MANAG, V28, P423, DOI 10.1016/J.IJHM.2008.11.005            21
KIM YG, 2010, INT J HOSP MANAG, V29, P216, DOI 10.1016/J.IJHM.2009.10.015            19
QUAN S, 2004, TOURISM MANAGE, V25, P297, DOI 10.1016/S0261-5177(03)00130-4           19
REALE D, 2007, BIOL REV, V82, P291, DOI 10.1111/J.1469-185X.2007.00010.X             19
CHANG RCY, 2010, ANN TOURISM RES, V37, P989, DOI 10.1016/J.ANNALS.2010.03.007        16
PLINER P, 2006, FRONT NUTR SCI, P75, DOI 10.1079/9780851990323.0075                  16
FORNELL C, 1981, J MARKETING RES, V18, P39, DOI 10.2307/3151312                      15
JI MJ, 2016, TOURISM MANAGE, V57, P387, DOI 10.1016/J.TOURMAN.2016.06.003            15
MAK AHN, 2012, INT J HOSP MANAG, V31, P928, DOI 10.1016/J.IJHM.2011.10.012           15
CHANG RCY, 2011, TOURISM MANAGE, V32, P307, DOI 10.1016/J.TOURMAN.2010.02.009        14
FISCHLER C, 1988, SOC SCI INFORM, V27, P275, DOI 10.1177/053901888027002005          14
GREENBERG RUSSELL, 2003, P175                                                        14
MARTIN LB, 2005, BEHAV ECOL, V16, P702, DOI 10.1093/BEHECO/ARI044                    14
SOL D, 2011, PLOS ONE, V6, DOI 10.1371/JOURNAL.PONE.0019535                          14

2: Co-citation analysis structure

Citation analysis is another remarkable tool of the bibliometric analysis offered by the package. It shows the structure of a specific field through the links between nodes (e.g. authors, articles, journal). The interesting option is that the edges can be interpreted differently depending on the type of network, i.e. co-citations, direct citations, bibliographic linking, etc…. This is very useful and can be exploited quite a lot.

Below we have taken the three standard examples shown in the original reference but adapted them to our database.

First, a co-citation network showing the relationships between the cited-referred works (nodes).

Second, a co-citation network that uses the cited journals as the unit of analysis.

The dimensions useful for commenting on co-citation networks are: (i) centrality and peripherality of nodes, (ii) their proximity and distance, (iii) strength of links, (iv) clusters, (v) bridging contributions.

Third, a historiography that is built on direct quotations. It traces the intellectual links in a historical order. The cited works of thousands of authors contained in a collection of published scientific articles are sufficient to reconstruct the historiographical structure of the field, pointing out the basic works in it.

wos=metaTagExtraction(wos,"CR_SO",sep=";")
NetMatrix <- biblioNetwork(wos, analysis = "co-citation", network = "sources", sep = ";")
net=networkPlot(NetMatrix, n = 50, Title = "Co-Citation Network", type = "auto", size.cex=TRUE, size=15, remove.multiple=FALSE, labelsize=1,edgesize = 10, edges.min=5)

Main statistics about the network

 Size                                  2876 
 Density                               0.034 
 Transitivity                          0.364 
 Diameter                              4 
 Degree Centralization                 0.481 
 Average path length                   2.284 
 

3: Historiography - Direct citation links

This analysis provides us with direct citation links by author and year, which is interesting to see who cites whom and when.

WOS DB:
Searching local citations (LCS) by reference items (SR) and DOIs...

Analyzing 10475 reference items...

Found 49 documents with no empty Local Citations (LCS)

 Legend

                                                                       Label
1              HEINRICH B, 1995, ANIM BEHAV DOI 10.1016/0003-3472(95)80130-8
2                        DAY RL, 2003, ANIM BEHAV DOI 10.1006/ANBE.2003.2074
3           METTKE-HOFMANN C, 2005, BEHAVIOUR DOI 10.1163/156853905774539427
4               METTKE-HOFMANN C, 2005, BEHAV ECOL DOI 10.1093/BEHECO/ARH159
5         ECHEVERRIA AI, 2008, ETHOLOGY DOI 10.1111/J.1439-0310.2008.01512.X
6              KIM YG, 2009, INT J HOSP MANAG DOI 10.1016/J.IJHM.2008.11.005
7      METTKE-HOFMANN C, 2009, ETHOLOGY DOI 10.1111/J.1439-0310.2009.01632.X
8                  BERGMAN TJ, 2009, ANIM COGN DOI 10.1007/S10071-008-0171-2
9              KIM YG, 2010, INT J HOSP MANAG DOI 10.1016/J.IJHM.2009.10.015
10               MORAND-FERRON J, 2011, BEHAV ECOL DOI 10.1093/BEHECO/ARR120
11              RAUDENBUSH B, 2012, APPETITE DOI 10.1016/J.APPET.2012.02.042
12             KIM YG, 2013, INT J HOSP MANAG DOI 10.1016/J.IJHM.2012.06.005
13 DIMITROVSKI D, 2017, J TRAVEL TOUR MARK DOI 10.1080/10548408.2016.1193100
14      MAK AHN, 2017, ASIA PAC J TOUR RES DOI 10.1080/10941665.2016.1175488
15    GRIFFIN AS, 2017, CURR OPIN BEHAV SCI DOI 10.1016/J.COBEHA.2017.01.004
16          HSU FC, 2018, TOUR RECREAT RES DOI 10.1080/02508281.2018.1475879
17                AKYUZ BG, 2019, ANATOLIA DOI 10.1080/13032917.2019.1595072
18         BAAH NG, 2020, CURR ISSUES TOUR DOI 10.1080/13683500.2019.1619676
19                   LAI MY, 2020, J TRAVEL RES DOI 10.1177/0047287519867144
20         OKUMUS B, 2021, TOUR MANAG PERSPECT DOI 10.1016/J.TMP.2020.100773
                                                                                                                            Author_Keywords
1                                                                                                                                      <NA>
2                                                                                                                                      <NA>
3                                                                             CONTEXT SPECIFICITY; EXPLORATION; NEOPHILIA; NEOPHOBIA; BIRDS
4                                                              ENVIRONMENTAL ASSESSMENT; NEOPHILIA; NOMADS; PARROTS; PSITTACIDAE; RESIDENTS
5                                                                                                                                      <NA>
6                                                 LOCAL FOOD; FOOD CONSUMPTION; FOOD TOURISM; CULINARY TOURISM; MOTIVATION; GROUNDED THEORY
7                                                                                                                                      <NA>
8                        NOVEL OBJECTS; NEOPHILIA; EXPLORATION; PRIMATE; NEOPHOBIA; THEROPITHECUS; GELADA; PAPIO URSINUS; BABOON; COGNITION
9                                                                          FOOD INVOLVEMENT; FOOD NEOPHOBIA; LOCAL FOOD; FESTIVALS; LOYALTY
10                                                       AGE DIFFERENCES; INNOVATION; PERSONALITY; PROBLEM SOLVING; SOCIAL; LEARNING; STATE
11                                                                                        PSYCHOLOGY; FOOD NEOPHOBIA; PHYSIOLOGICAL AROUSAL
12                                                                       EMPIRICAL VERIFICATION; LOCAL FOOD; FOOD CONSUMPTION; FOOD TOURISM
13                       FOOD MARKETS; SATISFACTION; TOURIST ATTRACTION; MOTIVATIONS; FOOD MARKET; LOVERS; FOOD NEOPHILIA; BARCELONA; SPAIN
14 TOURIST FOOD CONSUMPTION; FOOD-RELATED PERSONALITY TRAITS; FOOD; NEOPHOBIA; VARIETY-SEEKING; TOURIST'S PARADOX; SENSORY-SPECIFIC SATIETY
15                                                                                                                                     <NA>
16                                                      FOOD TOURISM; FOOD NEOPHOBIA; SENSATION-SEEKING; TRADITIONAL FOOD CHOICE; BEHAVIOUR
17                                                  CULINARY TOURISM; GASTRONOMY TOURISM; LOCAL FOOD; TRAVEL MOTIVATIONS; TRAVEL INTENTIONS
18                                                               GASTRONOMY TOURISM; FOOD TOURISM; NEOPHOBIA; NEOPHILIC TENDENCY; ATTITUDES
19                              DESTINATION FOOD IMAGE; FOOD TOURISM; FOOD NEOPHOBIA; POTENTIAL CHINESE; TOURIST; FORMATIVE IMAGE CONSTRUCT
20                                        GENDER; GENERATION; FOOD NEOPHOBIA; FOOD NEOPHILIA; FOOD TOURISM; ETHNIC; FOOD; CHINESE TRAVELERS
                                                                                                                                                            KeywordsPlus
1                                                                                                                                                                   <NA>
2                       TAMARINS LEONTOPITHECUS-ROSALIA; PRIMATE COGNITION; FEEDING NEOPHOBIA; CALLITHRIX-KUHLI; TOOL USE; BEHAVIOR; EVOLUTION; PATTERNS; SAGUINUS; SIZE
3         GUPPY POECILIA-RETICULATA; INDIVIDUAL-DIFFERENCES; EXPLORATORY-BEHAVIOR; GREAT TITS; REALIZED HERITABILITY; RISK-TAKING; PARUS-MAJOR; ANIMALS; PIGS; NEOPHOBIA
4                                       NUCLEOTIDE-SEQUENCES; DECISION-MAKING; BIRD COMMUNITY; PHYLOGENY; INFORMATION; STRATEGIES; NEOPHOBIA; BEHAVIOR; STIMULUS; GUINEA
5                                                                                  HOUSE SPARROWS; GROUP-SIZE; NEOPHOBIA; EXPLORATION; RAVENS; NEOPHILIA; BEHAVIOR; SONG
6                                                                                                      TOURISM; CHOICE; PERCEPTION; SELECTION; PATTERNS; HERITAGE; TASTE
7                                                       REALIZED HERITABILITY; MEMORY; EXPLORATION; BEHAVIOR; INFORMATION; HIPPOCAMPUS; FAMILIARITY; SUCCESS; LIFE; CUES
8                                         BEHAVIORAL ECOLOGY; FEEDING NEOPHOBIA; GREAT TITS; TRADE-OFFS; EVOLUTION; MONKEYS; EXPLORATION; HABITAT; BIRDS; RESPONSIVENESS
9                                                         CONSUMER INVOLVEMENT; NEOPHOBIA; TOURISM; CHOICE; SCALE; FAMILIAR; UNFAMILIAR; ATTITUDES; VARIABLES; SELECTION
10                  TIT PARUS-MAJOR; BEHAVIORAL FLEXIBILITY; INDIVIDUAL VARIATION; FORAGING; INNOVATION; SOCIAL RANK; GREAT TITS; BODY-SIZE; NEOPHOBIA; NEOPHILIA; BIRDS
11                                                                                                                                                                  <NA>
12                                                                                                                                          SCALE DEVELOPMENT; NEOPHOBIA
13                                       STRUCTURAL MODEL; CULINARY TOURISM; FARMERS MARKETS; NEOPHOBIA; FESTIVAL; ATTITUDES; VISITORS; SCALE; SATISFACTION; DESTINATION
14                                                              HONG-KONG; NEOPHOBIA; SCALE; CHOICE; MODEL; ANTECEDENTS; INFORMATION; EXPERIENCES; ACCEPTANCE; ATTITUDES
15       INVASION SUCCESS; BEHAVIORAL FLEXIBILITY; REPRODUCTIVE OUTPUT; PERSONALITY-TRAITS; TROPHIC DYNAMICS; NEST PREDATION; RISK-TAKING; LAND-COVER; BRAIN SIZE; BIRDS
16                                 SENSATION SEEKING; PLANNED BEHAVIOR; LOCAL FOOD; PERSONALITY-TRAITS; EUROPEAN FOOD; TOURISM; ATTITUDES; CHOICE; NEOPHOBIA; ACCEPTANCE
17                                  CULINARY TOURISM; CONSUMER INVOLVEMENT; DESTINATION; SEGMENTATION; NEOPHOBIA; SCALE; NATIONALITY; INFORMATION; ATTRACTION; NUTRITION
18                                             PURCHASE INTENTIONS; CONSUMER ATTITUDES; NOVELTY-SEEKING; HONG-KONG; FOOD; BEHAVIOR; ANTECEDENTS; GASTRONOMY; CHOICE; FIT
19 PARTIAL LEAST-SQUARES; STRUCTURAL EQUATION MODELS; LOCAL FOOD; CULINARY; TOURISM; FORMATIVE MEASUREMENT; PSYCHOLOGICAL-FACTORS; PLS; SATISFACTION; INFORMATION; PLACE
20                                        EATING BEHAVIOR; NOVELTY-SEEKING; FAMILIAR; CHILDREN; HABITS; SATISFACTION; WILLINGNESS; RESEMBLANCE; PERSONALITY; INVOLVEMENT
                                DOI Year LCS GCS
1      10.1016/0003-3472(95)80130-8 1995  10  71
2            10.1006/anbe.2003.2074 2003   8 103
3        10.1163/156853905774539427 2005   7  54
4             10.1093/beheco/arh159 2005   7  42
5  10.1111/j.1439-0310.2008.01512.x 2008   6  44
6        10.1016/j.ijhm.2008.11.005 2009  21 346
7  10.1111/j.1439-0310.2009.01632.x 2009   5  33
8         10.1007/s10071-008-0171-2 2009   5  58
9        10.1016/j.ijhm.2009.10.015 2010  19 178
10            10.1093/beheco/arr120 2011   5 106
11      10.1016/j.appet.2012.02.042 2012   5  41
12       10.1016/j.ijhm.2012.06.005 2013   8  85
13    10.1080/10548408.2016.1193100 2017   8  40
14    10.1080/10941665.2016.1175488 2017  13  73
15     10.1016/j.cobeha.2017.01.004 2017   6  51
16    10.1080/02508281.2018.1475879 2018   6  21
17    10.1080/13032917.2019.1595072 2019   4  10
18    10.1080/13683500.2019.1619676 2020   5  13
19         10.1177/0047287519867144 2020   6  34
20        10.1016/j.tmp.2020.100773 2021   5  11

4: Conceptual structure - Co-word analysis

Co-Word Networks show the conceptual structure, which uncovers the links between concepts through co-occurrences of terms.

The conceptual structure can be used to understand the topics being discussed (research front) and to identify which are the most important and most recent topics.

The tool divides the whole time span into different periods and compares the conceptual structures which is useful to analyze the evolution of the topics over time.

The package is able to analyze keywords, but also terms in article titles and abstracts. It does this by means of network analysis or correspondence analysis (CA) or multiple correspondence analysis (MCA). CA and MCA visualize the conceptual structure in a two-dimensional graph, which I also show below.

NetMatrix <- biblioNetwork(wos, analysis = "co-occurrences", network = "keywords", sep = ";")
net=networkPlot(NetMatrix, normalize="association", n = 50, Title = "Keyword Co-occurrences", type = "fruchterman", size.cex=TRUE, size=20, remove.multiple=F, edgesize = 10, labelsize=5,label.cex=TRUE,label.n=30,edges.min=2)

suppressWarnings(
CS <- conceptualStructure(wos, method="MCA", field="ID", minDegree=15, clust=5, stemming=FALSE, labelsize=15,documents=20)
)

5: Thematic maps

The co-word analysis draws clusters of the keywords. Themes are considered, whose density and centrality can be used to classify the themes and draw a two-dimensional diagram.

The thematic map is a very intuitive diagram and we can analyze the topics according to the quadrant in which they are placed: (1) upper right quadrant: main topics (as qualified by the authors of the tool); (2) lower right quadrant: basic topics; (3) lower left quadrant: emerging or disappearing topics; (4) upper left quadrant: very specialized/niche topics.

El análisis de co-palabras dibuja clusters de las palabras clave. Se consideran temas, cuya densidad y centralidad pueden utilizarse para clasificar los temas y trazar un diagrama bidimensional.

Map=thematicMap(wos, field = "ID", n = 250, minfreq = 4,
  stemming = FALSE, size = 0.7, n.labels=5, repel = TRUE)
plot(Map$map)

The description of the cluster can then be requested from the program:

Clusters=Map$words[order(Map$words$Cluster,-Map$words$Occurrences),]
library(dplyr)

Attaching package: 'dplyr'

The following objects are masked from 'package:stats':

    filter, lag

The following objects are masked from 'package:base':

    intersect, setdiff, setequal, union

CL <- Clusters %>% group_by(.data$Cluster_Label) %>% top_n(5, .data$Occurrences)
CL

# A tibble: 24 × 9
# Groups:   Cluster_Label [4]
   Occurrences Words           Cluster Color    Cluster_Label Cluster_Frequency btw_centrality clos_centrality pagerank_centrality
         <dbl> <chr>             <dbl> <chr>    <chr>                     <dbl>          <dbl>           <dbl>               <dbl>
 1           9 willingness           1 #E41A1C… willingness                 106           579.         0.00193             0.0112 
 2           7 acceptance            1 #E41A1C… willingness                 106           565.         0.00189             0.00872
 3           6 familiar              1 #E41A1C… willingness                 106           159.         0.00176             0.00660
 4           5 eating behavior       1 #E41A1C… willingness                 106           334.         0.00180             0.00572
 5           5 food neophobia        1 #E41A1C… willingness                 106           226.         0.00181             0.00480
 6          48 neophobia             2 #377EB8… neophobia                   416          4804.         0.00242             0.0438 
 7          29 behavior              2 #377EB8… neophobia                   416          4508.         0.00239             0.0271 
 8          21 neophilia             2 #377EB8… neophobia                   416          1163.         0.00205             0.0219 
 9          13 evolution             2 #377EB8… neophobia                   416          1001.         0.00204             0.0130 
10          12 exploration           2 #377EB8… neophobia                   416           936.         0.00204             0.0136 
# ℹ 14 more rows

6: Social structure - Collaboration analysis

This last section is also interesting. Collaborative networks show how authors, institutions (e.g., universities or departments) and countries relate to others in the field we are analyzing, in this case neophilia/neophobia in tourism. For example, the first figure below is a “Co-author network”. It uncovers regular study groups, hidden groups of scholars and key authors. The second figure is called an “Educational Collaboration Network” and uncovers relevant institutions in a specific research field and their relationships.

NetMatrix <- biblioNetwork(wos, analysis = "collaboration",  network = "authors", sep = ";")
net=networkPlot(NetMatrix,  n = 50, Title = "Author collaboration",type = "auto", size=10,size.cex=T,edgesize = 3,labelsize=1)

NetMatrix <- biblioNetwork(wos, analysis = "collaboration",  network = "universities", sep = ";")
net=networkPlot(NetMatrix,  n = 50, Title = "Edu collaboration",type = "auto", size=4,size.cex=F,edgesize = 3,labelsize=1)

wos <- metaTagExtraction(wos, Field = "AU_CO", sep = ";")
NetMatrix <- biblioNetwork(wos, analysis = "collaboration",  network = "countries", sep = ";")
net=networkPlot(NetMatrix,  n = dim(NetMatrix)[1], Title = "Country collaboration",type = "circle", size=10,size.cex=T,edgesize = 1,labelsize=0.6, cluster="none")

Final comments

Apart from this library that works in R, the authors have created an application that does the same without programming, and there is even an additional tool, also useful. The only thing you need is to have the DB in the correct format (merged WoS and Scopus, if you want, etc…) because otherwise, it does not read it but if you do it right, the result is the same as we have shown here but without programming R practically (only to transform the files).

References

This work is done with the R package bibliometrix and on the basis proposed by Aria and Cuccurullo, adapted and modified for our research line.

Aria, M., & Cuccurullo, C. (2017). bibliometrix: An R-tool for comprehensive science mapping analysis, Journal of Informetrics, 11(4), pp 959-9753643, (https://www.bibliometrix.org).

2026

September 29, 2026

Mapping the Miles: A Running Log as a Calendar Heatmap

heatmap

data viz

running

Running heatmap

No matching items

2023

December 23, 2023

Understanding Linear Regression

linear regression

correlation

goodness of fit

linear regression

No matching items

2022

September 9, 2022

Scientific Maps Analysis with the `R` package `bibliometrix`

bibliometry

mapping

bibliometric analysis

Example of scientific maps for bibliometric analysis

10.5281/zenodo.10079724

No matching items

2021