Trends in the Brazilian Labor Market Based on Socioeconomic Indicators of Occupation: Bi

Alice Juliane da Costa Rodrigues; Fernando Celso de Campos

DOI: 10.22167/2675-6528-202602631

Article derived from a Course Conclusion Work (TCC), with content based on the student’s original work and adapted to the editorial format of the E&S Magazine with the support of the ResumeAI tool, an artificial intelligence solution developed by Instituto Pecege for textual synthesis and organization.

Summary

The analysis of socioeconomic indicators proved fundamental to understanding the dynamics of the Brazilian labor market, historically marked by racial, regional, and economic inequalities. With the objective of simulating the analysis of these phenomena and supporting the development of analytical skills, an interactive dashboard was developed in MS Power BI®. The adopted methodology was descriptive and quantitative, based on a fictitious database, structured from characteristics and variables inspired by the National Household Sample Survey (PNAD Contínua), covering the period from 2022 to 2024. Key indicators were defined, such as employment and unemployment rates, average income, and demographic distributions. The dashboard allowed for interactive exploration of these indicators, facilitating the identification of patterns, trends, and inequalities. The simulated results reproduced behaviors consistent with the literature and real databases, evidencing variations in employment and income levels among population groups, as well as structural disparities influenced by schooling and gender. It was concluded that the application demonstrated potential as a tool to support analysis and decision-making for academic and data science learning purposes.

Keywords: Business data; MS Power BI®; Data visualization.

1. Introduction

The analysis of socioeconomic indicators is fundamental for understanding the dynamics of the Brazilian labor market, a scenario historically marked by deep inequalities. The legacy of the colonial period, which prioritized the extraction of wealth over national development, left marks that persist in the formation of the republican country. These disparities are manifested in contemporary times, especially in racial and regional dimensions (Nogueira and Souza, 2017).

The formal abolition of slavery, in 1888, was not accompanied by structured inclusion policies that promoted the integration of the freed population into society. The lack of access to employment, housing, and education resulted in the occupation of peripheral areas, which evolved into current favelas. Such historical problems still contribute to the maintenance of social and economic inequalities, with many of these regions lacking essential public policies, such as sanitation, health, and education (Schwarcz and Starling, 2015).

These inequalities are corroborated by recent data. In 2022, the white population had an income 87% higher than the black population, according to the Institute of Applied Economic Research (IPEA). The same survey reveals that black men and women make up approximately 80% of the poorest ten percent in Brazil. Regionally, the Regional Inequality Map (Grin et al., 2021) indicates that most of the national Gross Domestic Product is concentrated in the South and Southeast regions, highlighting structural characteristics of economic development and challenges for effective public policies.

Faced with this complex scenario, data visualization emerges as a strategic resource for understanding the trends of the Brazilian labor market and for subsidizing the formulation of public policies. By transforming large volumes of data into clear, accessible, and interpretable information, visualization contributes to broadening the understanding of social and economic phenomena, strengthening evidence-based decision-making (Concannon et al., 2019).

In this context, data visualization tools, such as MS Power BI, offer resources for the efficient analysis of large volumes of information. They allow for the monitoring of indicators through interactive dashboards, the identification of trends, and support for strategic planning (Becker and Gould, 2019). Power BI, recognized as a Business Intelligence tool (Microsoft, 2025), enables the integration, modeling, and visualization of data from multiple sources, converting them into easily interpretable interactive reports and dashboards.

Power BI’s capability is expanded by the use of languages such as Data Analysis Expressions (DAX) for creating analytical measures and the M language, used in Power Query for data extraction, transformation, and loading (ETL) processes. These features allow for the construction of complex analyses and advanced data preparation, even in contexts with large volumes of information (Microsoft, n.d.). Thus, the tool contributes to the democratization of data use and the strengthening of an evidence-based culture in organizations (Davenport and Patil, 2012).

Although widely used official databases exist, such as those provided by the Brazilian Institute of Geography and Statistics (IBGE) through the Continuous National Household Sample Survey (PNAD Contínua), this work opts for a fictitious database. This database was structured from characteristics and variables inspired by PNAD Contínua, with the objective of simulating scenarios and enabling the practical application of data analysis and visualization techniques.

The choice of simulated data is justified by the need to provide a controlled environment for the development of analytical skills and the exploration of plausible scenarios, promoting an integrated and evidence-based analysis, useful for academic and data analysis learning purposes. Thus, the objective of this work is to develop a dashboard that explores indicators of the Brazilian labor market, based on a simulated database, allowing the identification of patterns and trends and contributing to the development of analyses aimed at supporting decision-making and formulating social inclusion strategies.

2. Material and Methods

This research was characterized as an applied study of a descriptive nature, with a quantitative approach. Its main objective was to develop an application focused on the visual presentation of Brazilian labor market trends, seeking to generate information that could support decision-making in various contexts. For this purpose, a fictitious database was used, built to simulate the structure and variables of the Continuous National Household Sample Survey (PNAD Contínua), from the Brazilian Institute of Geography and Statistics (IBGE).

The adopted methodology was structured in six sequential steps, aiming at the construction of the analytical dashboard. Firstly, data collection was carried out, followed by processing, modeling, and analysis. Subsequently, visualization aspects were defined, the dashboard development proceeded, and finally, tests and solution implementation were performed. This process flow guided the construction of the work, from the definition of parameters to the elaboration of the analytical dashboard.

In the first stage, data collection, the methodological parameters and references were defined based on public information from IBGE, especially from PNAD Contínua. From these references, a synthetic database was built, generated with the support of artificial intelligence, with the purpose of simulating the behavior and structure of the original information. The temporal scope adopted for the simulation covered the period from 2022 to 2024, aligned with the analytical objectives of the study.

The main axes of analysis were defined as the dynamics of the labor market, the educational profile of the employed population, technical qualification, and formal labor income. Additionally, analytical filters were established for segmentation by gender, schooling, year, location, and age group, which structured the panel’s interactivity. These elements were essential for building an analytical environment that allowed for the exploration of plausible scenarios.

The second stage consisted of data processing, with the aim of ensuring its consistency and reliability. This process included standardizing formats and structures, removing duplicates, and handling missing values (Redman, 2001). Variables such as dates and monetary values were also standardized, and temporal aggregations were performed at different granularity levels, in addition to manual sample checks to validate the quality of the applied transformations (Kimball and Ross, 2013).

In the third stage, data modeling and analysis, with the data duly prepared, the analytical environment structuring began. The modeling followed the dimensional paradigm, organizing the information into a central fact table, named `base_dashboard_pnad`, and multiple dimension tables, such as gender, race, schooling, age group, federative unit, and income (Kimball and Ross, 2013).

The relationships between the tables were defined in a one-to-many (1:N) format, ensuring the correct propagation of filters in the MS Power BI® environment and the consistency of analyses. Measures were developed using the Data Analysis Expressions (DAX) language, enabling the construction of derived metrics, comparative analyses, and dynamic cross-referencing. Exploratory analysis employed descriptive statistics techniques and time series analysis, according to Chatfield (2004).

Among the main indicators constructed, the unemployment rate, the employment rate, the participation rate, and the average income stood out. Additionally, indicators related to the underutilization of the labor force, such as discouragement and underemployment, were developed. These indicators were fundamental to representing the dynamics of the Brazilian labor market in the panel.

The fourth stage focused on the definition and modeling of visualization aspects, seeking to ensure clarity, accessibility, and efficiency in data communication. The construction of visualizations was guided by principles of visual perception and best practices of “data storytelling”, prioritizing information readability, comparability between categories, and reduction of visual noise (Knaflic, 2015; Few, 2013).

For temporal analysis, line graphs were used, while for comparisons between groups, such as educational categories, gender, and federative units, bar graphs were chosen. Interactive elements, such as filters and data segmentations, were incorporated, allowing the user to dynamically explore different analytical cuts, such as year, age group, education level, gender, and location.

Visual standardization was a central aspect, with the definition of consistent patterns of colors, scales, and legends to facilitate the interpretation of indicators and reduce ambiguities (Cairo, 2016). Specific colors were associated with types of indicators, with shades of growth and stability for positive aspects, and alert colors for critical indicators, such as vacancy and underutilization.

In the fifth stage, dashboard development, with the structured database and defined visual elements, the interactive “dashboard” was developed in MS Power BI®. The measures created in DAX were integrated into the interface, allowing for dynamic data exploration through filters and segmentation. The dashboard was structured into thematic tabs, organized according to the analytical axes defined in the requirements gathering stage.

Thematic tabs covered the dynamics of the labor market, the educational profile of the employed population, technical qualification, and labor income. This organization allowed the user to navigate between different analytical perspectives, maintaining visual and logical coherence in data exploration, providing an integrated and interactive view of the main dynamics of the Brazilian labor market.

The sixth and final stage involved the testing and implementation of the dashboard, with the objective of validating data consistency, indicator reliability, and solution usability. Validation tests of the calculations implemented in DAX were performed, comparing the dashboard results with the values from the database. The consistency of interactive filters was also tested, evaluating the behavior of segmentations by gender, education level, age group, location, and period.

The usability of the dashboard was evaluated qualitatively, considering the clarity of information, visual organization, readability of graphs, and ease of navigation. After the tests, minor adjustments were made related to the formatting of indicators, organization of visual elements, and refinement of interactions. The dashboard was consolidated as a functional application for socioeconomic data analysis, suitable for exploratory and educational purposes.

3. Results and Discussion

Based on the previously defined methodological steps, the obtained results are presented, analyzed, and interpreted, respecting the logical sequence of the project development and grounded in good practices of data modeling and information visualization. Initially, a requirements survey was conducted to structure an analytical dashboard aimed at understanding the dynamics of the Brazilian labor market, using simulated data inspired by the Continuous National Household Sample Survey (PNAD Contínua). This process allowed for the definition of the analysis axes and the interactive filters that would guide data exploration.

The main axes of analysis defined were the dynamics of the labor market, the educational profile of the employed population, technical qualification, and formal labor income. Additionally, analytical filters were established by gender, schooling, year, location, and age group, which structure the panel’s interactivity. The data collection stage consisted of building a synthetic database, based on parameters and references from PNAD Contínua, covering the period from 2022 to 2024. This database resulted in a set of structured data, containing relevant socioeconomic variables such as occupation, unemployment, income, and demographic characteristics.

This simulated database allowed the exploration of different labor market scenarios, preserving structural patterns observed in real data, such as regional inequalities and variations in income and occupation levels. Although the use of fictitious data limits direct inference about empirical reality, this approach proved adequate for the study’s objective, enabling the development and application of analytical and visualization techniques in a controlled environment. The construction of dashboards, as highlighted by Few (2013), should focus on the user and clear analytical objectives to avoid informational overload and favor decision-making.

In the data processing stage, the procedures applied aimed to ensure the consistency, standardization, and analytical reliability of the information. This included checking data types, standardizing variables such as dates and monetary values, and performing sample validations to guarantee the coherence of the records. No relevant occurrences of duplicates or missing values requiring more complex treatments were identified, which is attributed to the synthetic nature of the database.

Additionally, structural adjustments were made to prepare the data for the analytical environment, ensuring its compatibility with subsequent modeling and analysis steps. These actions were crucial to ensure data integrity and the reliability of the constructed indicators. According to Redman (2001), data quality is a determining factor for generating reliable information, and the processing of data is essential to reduce inconsistencies and improve the accuracy of analyses.

The treatment stage, although efficient in ensuring the minimum necessary data quality, presented low complexity in interventions. This is due to the use of a synthetic base, which reduces the representativeness of this phase when compared to real contexts, where problems such as missing data, structural inconsistencies, and informational noise are more frequent and require greater technical effort. However, the adopted process was sufficient to ensure the reliability of the data used in the subsequent stages of the study.

With the data duly prepared, the modeling and analysis stage began, with the objective of structuring the analytical environment and enabling the construction of indicators representative of the dynamics of the Brazilian labor market. The modeling was carried out based on the dimensional paradigm, organizing the information into a central fact table, named `base_dashboard_pnad`, and multiple dimension tables. The dimension tables included gender, race, education level, age group, federal unit, and income, allowing for segmentation and analysis from different perspectives.

The relationships between the tables were defined in a one-to-many (1:N) format, which ensured the correct propagation of filters in the MS Power BI® environment and guaranteed the consistency of the analyses. This type of modeling favors both query performance and end-users’ understanding of the data, as highlighted by Kimball and Ross (2013). Based on this structure, measures were developed using the Data Analysis Expressions (DAX) language, enabling the creation of analytical indicators and derived metrics.

Among the main indicators constructed with DAX, the unemployment rate, employment rate, participation rate, average income, as well as indicators related to the underutilization of the labor force, such as discouragement and underemployment, stand out. The DAX measures allowed for dynamic and comparative analyses, including temporal variations, aggregations by different levels of granularity, and cross-tabulations between socioeconomic variables, expanding the exploratory capacity of the panel and enabling interactive and personalized analyses.

Within the scope of exploratory analysis, it was observed that the indicators showed variations throughout the analyzed period (2022–2024), evidencing differences between population groups. Variations were identified in occupation and income levels when considering breakdowns such as schooling, age group, and geographical location, reflecting structural inequalities in the labor market. The temporal analysis of the indicators allowed for the identification of oscillating behaviors over the periods, suggesting the influence of economic and social factors on the dynamics of employment and income, according to the importance of time series analysis for identifying patterns and trends (Chatfield, 2004).

Dimensional modeling combined with the use of Data Analysis Expressions (DAX) proved effective in structuring the analytical environment and enabling the construction of relevant indicators. The adopted approach allowed for flexibility in data exploration, favoring multidimensional and comparative analyses. However, as it is a synthetic database, the results obtained should be interpreted as behavioral simulations, not directly reflecting empirical reality. Nevertheless, the constructed structure demonstrates adherence to good data modeling and analysis practices, being fully applicable to real-world scenarios with official data.

In the stage of definition and modeling of visualization aspects, the aim was to represent the data clearly, accessibly, and pedagogically appropriately, guided by principles of visual perception and good “data storytelling” practices. For temporal analysis, line graphs were used, capable of highlighting trends and variations. For comparisons between groups, such as educational attainment categories, gender, and federative units, bar graphs were chosen, which facilitate the reading and ordering of values.

Elements of interactivity, such as filters and data segmentations, were incorporated, allowing the user to dynamically explore different analytical cuts, such as year, age group, schooling, gender, and location. This approach expanded the panel’s analysis capacity, enabling the construction of different perspectives from the same dataset. Visual standardization was a central aspect, with the definition of consistent standards for colors, scales, and legends, to facilitate the interpretation of indicators and reduce ambiguities.

According to Knaflic (2015), the adequate choice of visual elements is essential for effective data communication, prioritizing clarity and the elimination of unnecessary elements. Complementarily, Few (2013) highlights that efficient dashboards should be designed with a user focus, presenting information in an organized manner and without cognitive overload. The definition of visualization aspects proved adequate to meet the study’s objectives, contributing to the construction of a clear, interactive, and easily interpretable dashboard. However, the effectiveness of visualizations directly depends on the user’s profile and their level of familiarity with analytical tools, which can influence data interpretation. Nevertheless, the choices made follow consolidated good practices in the literature, favoring the communication of results in an accessible and structured way.

In the dashboard development stage, with the structured database, the defined indicators, and the visualization aspects modeled, an interactive dashboard was developed in MS Power BI®, with the objective of presenting the main dynamics of the Brazilian labor market in an integrated manner. The dashboard was structured into a segmented view by thematic tabs, organized according to the analytical axes defined in the requirements gathering stage, covering the dynamics of the labor market, educational profile, technical qualification, and work income.

This organization allowed the user to navigate between different analytical perspectives, maintaining visual and logical coherence in data exploration. In addition to the tab structure, the panel was designed to be highly interactive, incorporating filters and segmentations that allow analysis by variables such as year, gender, age group, education level, and geographic location. This interactivity enabled both an overview of the indicators and more specific analyses, according to the user’s interest.

The first section of the dashboard, focused on the labor market dynamics, presented indicators such as working-age population, participation rate, employment rate, unemployment, discouragement, and underutilization of the labor force. The evolution of the working-age population revealed regional and temporal variations throughout the period from 2022 to 2024. The overall unemployment rate and the participation rate also showed oscillations and significant differences between demographic groups, such as age groups and races, highlighting the complexity of the labor market.

The distribution of the workforce by gender and federative unit indicated regional and gender disparities in labor market insertion. The evolution of the unemployed, discouraged, and underemployed workforce over the quarters from 2022 to 2024 showed fluctuations, with peaks and troughs suggesting the influence of seasonal or economic factors. The analysis of these indicators revealed variations over the period, suggesting changes in labor market dynamics, and reinforcing the importance of segmented analyses to understand oscillations in participation and employment levels, as well as differences between population groups.

The second section of the survey, dedicated to the educational profile of the employed population, allowed for observing the distribution of individuals according to schooling levels. The results indicated that schooling levels directly influence labor market insertion, reflecting structural inequalities in access to opportunities. The distribution of the employed population by schooling and economic activity, as well as the income differential between women and men by schooling and type of employment, and between white and black individuals, highlighted the persistence of significant gaps.

The third section addressed aspects related to the technical qualification of the population, allowing an analysis of the relationship between education and insertion into the labor market. The visualizations in this section sought to highlight how different qualification levels impact indicators such as occupation and income, through comparative graphs. The analysis suggests that higher qualification levels are associated with better labor market insertion conditions, although this relationship is not uniform across all groups, indicating the influence of other structural factors.

The last section of the panel presented indicators related to labor income, focusing on the analysis of income distribution among different population groups. Bar charts and segmented comparisons were used to highlight income differences by education level, gender, and location. The results showed significant disparities in income levels, reinforcing the existence of structural inequalities in the labor market, especially when considering factors such as education level and gender. The evolution of real formal labor income and its variation by groups revealed the persistence of these inequalities.

From the integrated analysis of the different sections of the dashboard, it was possible to highlight the existence of temporal variations in labor market indicators throughout the analyzed period, as well as relevant differences in employment and participation levels among population groups. The direct influence of schooling on insertion and permanence in the labor market and the association between technical qualification and better employment conditions were evident. Furthermore, the presence of structural inequalities in income levels, especially between different educational levels and social groups, was a consistent finding.

The development of the dashboard in MS Power BI® allowed the consolidation of different dimensions of the labor market in a single environment, providing an integrated and interactive view of the data. The structure with thematic tabs contributed to the organization of information, facilitating navigation and interpretation of results. The interactivity of the dashboard, enabled by filters and slicers, expanded the analytical capacity of the tool, allowing the exploration of different scenarios from the same dataset.

However, as it is a synthetic database, the results should be interpreted with caution, being more suitable for analytical and educational purposes than for direct inferences about reality. Furthermore, the absence of external variables limits the understanding of macroeconomic factors that influence the labor market. Nevertheless, the developed panel demonstrates the potential of business intelligence tools in the analysis of socioeconomic data, highlighting their applicability in real contexts.

A relevant aspect in the development of the dashboard was the standardization of the color palette and the visual consistency of the indicators. We sought to establish a chromatic standard that would allow the user to quickly identify the nature of each metric, reducing ambiguities and facilitating the interpretation of information. Specific colors were associated with types of indicators: shades of growth and stability for positive aspects (occupancy, yield), and alert colors for critical indicators (unemployment, underutilization).

This standardization contributed to the creation of an intuitive visual language, allowing the user to recognize patterns and relationships more quickly. The consistency in applying colors to similar categories across different visualizations avoided the reinterpretation of legends and reduced cognitive load during analysis, which is crucial in interactive dashboards. According to Cairo (2016), the consistent use of visual elements is fundamental for building effective graphical representations, contributing to information organization and rapid data comprehension.

The definition of a consistent color palette proved effective in enhancing the readability and interpretation of indicators. However, color perception can vary among users, especially in cases of color blindness or different viewing contexts, which may partially impact the effectiveness of visual communication. Nevertheless, the adopted strategy follows good information design practices, contributing to a more intuitive and structured analytical experience.

After the development of the dashboard, tests were carried out with the objective of validating data consistency, indicator reliability, and solution usability. These tests were fundamental to ensure that the dashboard met the defined analytical objectives and provided an adequate end-user experience. Initially, validation tests of the calculations implemented in Data Analysis Expressions (DAX) were conducted, comparing the results presented in the dashboard with the values obtained directly from the database, verifying the correct application of measures and the precision of the indicators.

Next, consistency tests of the interactive filters were performed, evaluating the behavior of segmentations by gender, schooling, age group, location, and period. These tests ensured that user interactions resulted in coherent updates aligned with the data model’s logic, without distortions or inconsistencies in the visualizations. Additionally, the usability of the dashboard was qualitatively evaluated, considering aspects such as clarity of information, visual organization, readability of graphs, and ease of navigation between different tabs, aiming to ensure that the dashboard was intuitive for users with varying levels of familiarity with analytical tools.

After conducting the tests, specific adjustments were made related to indicator formatting, organization of visual elements, and refinement of interactions, contributing to the improvement of the user experience and the robustness of the solution. The testing phase proved essential to ensure the reliability and quality of the developed solution, highlighting the importance of validation in data analysis projects. However, as it is an application based on synthetic data and without implementation in a production environment, it was not possible to evaluate aspects such as large-scale performance, simultaneous use by multiple users, or integration with external real-time sources. Nevertheless, the solution demonstrates potential for practical applicability, especially in educational and analytical simulation contexts.

A comparative analysis between the scenario before and after the implementation of the business intelligence solution revealed significant evolution. Before the application, analytical productivity was characterized by manual and time-consuming processes, with the use of isolated spreadsheets for data consolidation. After the application, visualizations became automated and updated in real-time through the dashboard, resulting in a significant reduction in the time required for analysis.

The response time for report generation and analytical demand fulfillment, which was previously time-consuming, has transformed into rapid responses obtained through interactive filters and dynamic dashboards, increasing agility in decision-making. Resource savings, which previously required significant operational effort and presented a higher risk of rework, now benefit from automated processes and reduced reliance on manual intervention, optimizing the use of human and technological resources.

The quality of information, which was initially decentralized, inconsistent, and unstandardized, evolved into a structured, consistent, and automatically updated base, increasing the reliability and accuracy of analyses. The strategic vision, previously focused on the operational level and struggling to identify patterns, became integrated and multidimensional, allowing for more comprehensive analyses and strengthening support for strategic planning. Finally, indicator monitoring, which was manual and had a low update frequency, became automated and updated as data is processed, improving governance and continuous monitoring of indicators.

In summary, the developed dashboard explored indicators of the Brazilian labor market from a simulated database, reproducing behaviors consistent with the literature and real databases. The tool allowed the identification of patterns, trends, and inequalities, highlighting variations in occupation and income levels among population groups, as well as structural disparities influenced by schooling and gender. The application demonstrated potential as a support tool for analysis and decision-making for academic and data science learning purposes, despite the limitations inherent in using synthetic data for direct inferences about empirical reality.

4. Conclusion

The study aimed to develop a dashboard that explored indicators of the Brazilian labor market, based on a simulated database, allowing the identification of patterns and trends and contributing to the development of analyses aimed at supporting decision-making and formulating social inclusion strategies. It was found that the construction of an interactive dashboard in MS Power BI®, using a fictitious database inspired by the National Continuous Household Sample Survey for the period from 2022 to 2024, enabled the analysis of key indicators such as employment and unemployment rates, average income, and demographic distributions. The simulated results reproduced behaviors consistent with the literature and real databases, evidencing variations in employment and income levels among population groups and structural disparities influenced by education and gender. Dimensional modeling and the use of Data Analysis Expressions (DAX) proved effective in structuring the analytical environment, and visual standardization contributed to the clarity and interpretation of the data. The main contribution lies in the potential of the application as a tool to support analysis and learning in data science, offering an integrated and interactive view of labor market dynamics.

The main limitation is that the use of a synthetic database restricts direct inferences about empirical reality. Furthermore, the absence of external variables limited the understanding of macroeconomic factors influencing the labor market, and the non-implementation in a production environment prevented the evaluation of aspects such as scalability and real-time multi-user use. For future studies, it is suggested to use real databases and integrate with external data sources, such as macroeconomic indicators and public policy information, to broaden the analytical capacity of the panel. It is also recommended to include new indicators and conduct usability tests with end-users for a more in-depth evaluation of the tool’s impact on decision-making.

Bibliographic References

Becker, L. T.; Gould, E.M. 2019. Microsoft Power BI: Extending Excel to Manipulate, Analyze, and Visualize Diverse Data, Serials Review 45(3): 184-188.

CAIRO, Alberto. The truthful art: data, charts, and maps for communication. Berkeley: New Riders, 2016.

Chatfield, C.. 2004. The Analysis of Time Series: An Introduction. 6th Edition, Chapman and Hall, London.

Concannon D.; Herbst K.; Manley E. 2019. Developing a Data Dashboard Framework for Population Health Surveillance: Widening Access to Clinical Trial Findings. JMIR Form Res. 3(2): 1-13.

DAVENPORT, Thomas H.; PATIL, D. J. 2012. Data Scientist: The Sexiest Job of the 21st Century. Harvard Business Review, Boston, out. 2012. Link disponível em: https://hbr.org/2012/10/data-scientist-the-sexiest-job-of-the-21st-century. Acesso em: 28 out. 2025.

FEW, Stephen. 2013. Information Dashboard Design: Displaying Data for At-a-Glance Monitoring. 2. ed. Burlingame, CA: Analytics Press.

GRIN, Eduardo José; PIMENTEL DOS SANTOS, Fernando Burgos; FERNANDES, Gustavo Andrey de Almeida Lopes; BRESCIANI, Luís Paulo. 2021. O mapa regional das múltiplas desigualdades e do desenvolvimento humano no Brasil. São Paulo: Fundação Getulio Vargas. Link Disponível em: https://eaesp.fgv.br/sites/eaesp.fgv.br/files/pesquisa-eaesp-files/arquivos/b1_-_d521976c-0e1d-2ab6-d285-cacde252c490.pdf?utm_source=chatgpt.com. Acesso em: 23 set. 2025.

Kimball, R.; Ross, M. 2013. The Data Warehouse Toolkit: The Definitive Guide to Dimensional Modeling. 3ed. John Wiley & Sons, Indianapolis, IN, EUA. Link disponível em: https://ia801609.us.archive.org/14/items/the-data-warehouse-toolkit-kimball/The%20Data%20Warehouse%20Toolkit%20-%20Kimball.pdf. Acesso em: 23 set. 2025.

Knaflic, Cole Nussbaumer. 2015. Storytelling with Data: A Data Visualization Guide for Business Professionals. Wiley, 2015.

MICROSOFT 2025. Power BI documentation. Disponível em: https://learn.microsoft.com/power-bi/. Acesso em: 28 out. 2025.

NOGUEIRA, Oracy; SOUZA, Jessé. 2017. A elite do atraso: da escravidão à Lava Jato. Rio de Janeiro: Leya.

Redman, T.C. 2001. Data Quality: The Field Guide. Digital Press, Burlington, MA, EUA. Link disponível em: https://books.google.com.br/books?hl=pt-BR&lr=&id=iE2qsxDyEEYC&oi=fnd&pg=PR11&dq=Redman,+T.+2001.+Data+Quality:+The+Field+Guide&ots=6mnT-dJYPG&sig=qZWOzruirk06PVKpp_IXCfTgFio#v=onepage&q=Redman%2C%20T.%202001.%20Data%20Quality%3A%20The%20Field%20Guide&f=false. Acesso em: 23 set. 2025.

SCHWARCZ, Lilia Moritz; STARLING, Heloisa Murgel. 2015. Brasil: uma biografia. São Paulo: Companhia das Letras.

Article originating from the Final Course Project of the Specialization in Data Science and Analytics of the MBA USP/Esalq

To learn more about the course, click here and access the MBX Academy platform