Create charts in ChatGPT in seconds

Charts are an essential tool in data analysis and information visualization. They allow to visually represent the relationship between different variables and facilitate the understanding of patterns, trends and comparisons. In addition, graphics can convey complex information clearly and concisely, making them a valuable communication tool.

In the context of learning Python, creating charts is a great way to apply and consolidate the knowledge gained. By using libraries such as Matplotlib, Seaborn or Plotly, we can generate bar charts, line charts, scatter charts and many other types of visualizations. These libraries offer a wide range of customization and configuration options, allowing us to adapt the graphics to our specific needs.

Charts in ChatGPT

The creation of graphs in ChatGPT gives us several advantages to learn Python. First, it allows us to practice the syntax and functions needed to generate graphics in the Python language. As we explore different types of charts and adjust their parameters, we are applying concepts learned in programming, such as list manipulation, indexing, loops, and control structures.

In addition, when creating graphs in ChatGPT, we are working with real or fictitious data, which allows us to develop data manipulation and analysis skills. We can apply data cleansing and transformation techniques before visualizing it, which helps us better understand how data is handled in the context of visualization.

On the other hand, the graphs generated in ChatGPT provide us with immediate visual feedback on the results of our code. We can look at the graphs in real time and check if the data is represented correctly and if it meets our visualization goals. This helps us identify potential errors and make adjustments to our code to improve the visual presentation of the data.

Creating data that ChatGPT provides us for charts

The distribution of tourist expenditure by destination country will be analysed. The dataset is not provided by ChatGPT itself. I have not checked whether they are correct data or not.

The following code is used:

import matplotlib.pyplot as plt
import numpy as np

# Data on tourism expenditure in countries
years = [2019, 2020, 2021, 2022]
countries = ['Spain', 'France', 'Italy', 'Germany', 'UK', 'Greece']
spain = [92.278, 19.723, 22.961, 24.520]
france = [55.674, 17.312, 24.751, 28.157]
italy = [45.150, 17.720, 23.532, 27.613]
germany = [38.861, 16.361, 21.181, 24.553]
uk = [41.861, 8.542, 14.852, 16.963]
greece = [18.198, 4.370, 7.712, 9.628]

# Configuración del gráfico
plt.figure(figsize=(10, 6))  # Tamaño del gráfico
bar_width = 0.1
opacity = 0.8
index = np.arange(len(years))

# Colores para cada país
colors = ['blue', 'orange', 'green', 'red', 'purple', 'brown']

# Dibujar las barras para cada país y año
for i, country in enumerate(countries):
    plt.bar(index + i * bar_width, eval(country.lower()), bar_width, alpha=opacity, color=colors[i], label=country)
    # Mostrar el valor encima de cada columna
    for j, value in enumerate(eval(country.lower())):
        plt.text(j + i * bar_width, value + 0.5, str(value), ha='center', va='bottom', fontsize=5)

plt.xlabel('Años')
plt.ylabel('Gasto Turístico (miles de millones de euros)')
plt.title('Gasto Turístico de países europeos (2019-2022)')
plt.xticks(index + bar_width * (len(countries) / 2 - 0.5), years)
plt.legend()

# Mostrar el gráfico
plt.show()

# Guardar el gráfico en formato PNG
plt.savefig('nombre_del_archivo.png', dpi=300, bbox_inches='tight')

The following graph is obtained:

Checking the percentage change from year to year

Now we’re going to ask ChatGPT to create a table showing the percentage change from year to year for each country and then a graph similar to the one above showing that percentage change. This code is used:

import matplotlib.pyplot as plt
import numpy as np

# Datos del gasto turístico de los países
years = [2019, 2020, 2021, 2022]
countries = ['Spain', 'France', 'Italy', 'Germany', 'UK', 'Greece']
spain = [92.278, 19.723, 22.961, 24.520]
france = [55.674, 17.312, 24.751, 28.157]
italy = [45.150, 17.720, 23.532, 27.613]
germany = [38.861, 16.361, 21.181, 24.553]
uk = [41.861, 8.542, 14.852, 16.963]
greece = [18.198, 4.370, 7.712, 9.628]

# Calcular el porcentaje de cambio de año a año para cada país
spain_pct_change = [(spain[i+1] - spain[i]) / spain[i] * 100 for i in range(len(spain)-1)]
france_pct_change = [(france[i+1] - france[i]) / france[i] * 100 for i in range(len(france)-1)]
italy_pct_change = [(italy[i+1] - italy[i]) / italy[i] * 100 for i in range(len(italy)-1)]
germany_pct_change = [(germany[i+1] - germany[i]) / germany[i] * 100 for i in range(len(germany)-1)]
uk_pct_change = [(uk[i+1] - uk[i]) / uk[i] * 100 for i in range(len(uk)-1)]
greece_pct_change = [(greece[i+1] - greece[i]) / greece[i] * 100 for i in range(len(greece)-1)]

# Crear una tabla que muestra el porcentaje de cambio de año a año
table_data = np.vstack((spain_pct_change, france_pct_change, italy_pct_change, germany_pct_change, uk_pct_change, greece_pct_change))

# Configuración del formato de la tabla
row_labels = ['Spain', 'France', 'Italy', 'Germany', 'UK', 'Greece']
col_labels = ['2019-2020', '2020-2021', '2021-2022']
table_title = 'Porcentaje de Cambio de Año a Año del Gasto Turístico'
cell_text = [['{:.2f}%'.format(value) for value in row] for row in table_data]
colors = plt.cm.get_cmap('Set2', len(countries))(np.arange(len(countries)))

# Configuración del gráfico de barras
plt.figure(figsize=(10, 6))  # Tamaño del gráfico
bar_width = 0.1  # Modifica el valor para ajustar el tamaño de las barras
opacity = 0.8
index = np.arange(len(years)-1)

# Dibujar las barras para cada país y año
for i, country in enumerate(countries):
    values = table_data[i]
    pos_values = [value if value >= 0 else 0 for value in values]
    neg_values = [value if value < 0 else 0 for value in values]
    plt.bar(index + i * bar_width, pos_values, bar_width, alpha=opacity, color=colors[i], label=country)
    plt.bar(index + i * bar_width, neg_values, bar_width, alpha=opacity, color=colors[i])
    # Mostrar el valor en cada barra
    for j, value in enumerate(values):
        if value >= 0:
            plt.text(index[j] + i * bar_width, value + 0.5, '{:.2f}%'.format(value), ha='center', va='bottom', fontsize=8)
        else:
            plt.text(index[j] + i * bar_width, value - 0.5, '{:.2f}%'.format(value), ha='center', va='top', fontsize=8)

plt.xlabel('Años')
plt.ylabel('Cambio Porcentual del Gasto Turístico')
plt.title('Cambio Porcentual de los Gastos Turísticos de países europeos (2019-2022)')
plt.xticks(index + bar_width * (len(countries) / 2 - 0.5), col_labels)
plt.legend()

# Guardar el gráfico en formato PNG
plt.savefig('porcentual.png', dpi=300, bbox_inches='tight')

# Mostrar el gráfico
plt.show()

The data comes from the following table:

Country	2019-2020	2020-2021	2021-2022
Spain	-78,6265	16,41738	6,789774
France	-68,9047	42,97019	13,76106
Italy	-60,753	32,7991	17,34234
Germany	-57,8987	29,4603	15,91993
UK	-79,5944	73,87029	14,21357
Greece	-75,9864	76,47597	24,8444

And they give rise to the following graph:

Summary

In short, creating charts in ChatGPT is a great way to learn Python and consolidate the concepts of programming and data analysis. It allows us to apply Python syntax and functions in the context of data visualization, develop data manipulation and analysis skills, and get immediate visual feedback on our results. In addition, by generating graphs, we can communicate and convey information effectively, which is essential in data analysis and informed decision making.

On many occasions the result we obtain is not correct. The most appropriate way to solve it is to tell ChatGPT exactly that it is wrong and he will only debug the code little by little. The result obtained in this article has been after several attempts, asking the right questions to get what you are looking for.