From Collection to Decision: Provenance and Fair in the Transparency of Federal Universities’ Spending

André Ferreira Zeferino; Lilian da Silveira

DOI: 10.22167/2675-6528-202602662

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 study investigated the budgetary transparency and dynamics of the federal universities UFRJ and UFRRJ between 2020 and 2024. The objective was to understand the impact of disruptive events, such as the COVID-19 pandemic, on the financial sustainability of these institutions, grounding the research in FAIR principles and the PROV-W3C provenance model to ensure the integrity of records. The methodology integrated data collection from the Transparency Portal, using web scraping and API consumption in Python, and employed time series analysis with decomposition into trend, seasonality, and noise. The results revealed high fiscal rigidity in both universities, with a predominance of mandatory expenses for personnel and social charges, which peaked in 2024. It was observed that the nominal growth in current expenses in the post-pandemic period did not represent a real expansion of services, but rather the pressure of fixed costs and ongoing contracts. Capital investments suffered severe contractions, indicating physical and technological obsolescence. The methodological comparison between manual collection and the API identified significant nominal discrepancies, although there was convergence in effective cash flows. It was concluded that the institutions operated under a permanent crisis management regime, sacrificing infrastructure modernization for immediate operational survival, which requires new financing policies to ensure future academic excellence.

Keywords: FAIR Principles; Data Provenance; Time Series; Public Transparency; Federal Universities.

1. Introduction

The evolution of new public governance in Brazil has demanded more efficient and results-oriented management. As highlighted by Reis et al. (2024), this modernization not only meets social demands but also drives the country’s economic and commercial relations with the international scene, consolidating transparency as an indispensable pillar of contemporary public administration. In this sense, transparency in public management is configured as a structuring element of democracy, as it enables society to monitor and oversee the application of public resources. In Brazil, the Transparency Portal has consolidated itself as one of the main instruments of social control by making detailed information available on the Union’s budgetary and financial execution.

Within the scope of federal universities, such as the Federal University of Rio de Janeiro (UFRJ) and the Federal Rural University of Rio de Janeiro (UFRRJ), clarity in expense management plays a strategic role, being directly related to the sustainability of teaching, research, and innovation activities. However, the mere availability of raw data does not guarantee the effectiveness of transparency.

For information to be useful and conducive to true knowledge discovery, it is essential that data adhere to the FAIR Principles (Findable, Accessible, Interoperable, Reusable). As established by Wilkinson et al. (2016), research digital objects must be Findable, Accessible, Interoperable, and Reusable, with a specific emphasis on the ability of machines and computational agents to automatically find and utilize this data.

In this sense, transparency evolves into the concept of data stewardship, where the quality of the publication depends on rich metadata that ensures the integrity of digital assets in the long term (Wilkinson et al., 2016). A critical component of this integrity is provenance, which describes the entities, activities, and agents involved in the production and manipulation of a resource (World Wide Web Consortium, 2013). According to the definitions of the W3C PROV standard, provenance allows for the evaluation of data reliability and quality by offering a historical record of how information was generated and processed. For this study, maintaining a detailed provenance record is essential for the collected data to be accurately cited and scientifically validated.

Although the literature on the combined application of FAIR principles and the PROV-W3C model to university transparency in Brazil is still emerging, similar approaches in the area of public data governance demonstrate the fragilities of traditional extraction methods. Research relying solely on unstructured web scraping or the isolated consumption of consolidated reports tends to suffer from the instability of web interfaces and the absence of provenance metadata, compromising reproducibility.

In contrast, the methodology proposed here raises the level of reliability and auditability by formally registering each data transformation step, according to the PROV-W3C standard. This structure aligns with the international practices recommended by Wilkinson et al. (2016), ensuring that budgetary data is auditable regarding its origin and processing history. From this perspective, the use of digital tools and automated collection methods emerges as a necessary condition for the integrity of these records.

This study advances in technical analysis by incorporating Time Series Analysis methods, allowing not only the grouping of spending profiles but also a deep understanding of budgetary dynamics between the years 2020 and 2024. The application of Time Series is justified by the need to identify long-term trends, seasonal commitment cycles, and, above all, the impact of disruptive events, such as the COVID-19 pandemic, on the fiscal rigidity of institutions. Through the decomposition of this chronological data, the aim is to diagnose whether the observed variations reflect real gains or inflationary losses, offering a longitudinal view that surpasses static analysis and contributes to the construction of predictive models on the financial sustainability of universities. In this context, this study investigated the transparency and budgetary dynamics of the federal universities UFRJ and UFRRJ between the years 2020 and 2024, based on FAIR principles and the PROV-W3C provenance model to ensure the integrity of the records.

2. Material and Methods

The methodology adopted in this study was characterized by a quantitative approach, of a descriptive and exploratory nature, with the objective of analyzing the transparency and budgetary dynamics of two federal higher education institutions. The research focused on the expenses of the Federal University of Rio de Janeiro (UFRJ) and the Federal Rural University of Rio de Janeiro (UFRRJ), covering the period from 2020 to 2024. To achieve the proposed objectives, descriptive statistics techniques were integrated with advanced Time Series Analysis methods.

The development and execution environment for the analyses was Google Colab, using the Python programming language, version 3.14. For data manipulation and processing, consolidated libraries in Data Science were employed, such as “requests”, “pandas”, “matplotlib”, “numpy”, and “seaborn”. These tools were essential for the organization, visualization, and pre-processing of the datasets.

For Time Series modeling, the use of the libraries “statsmodels” and “prophet” was incorporated, which allowed the decomposition of the data into its trend, seasonality, and noise components. Additionally, the “K-Means” algorithm was applied via “scikit-learn” complementarily, aiming to validate the similarity between the identified temporal cycles. The visualization of provenance flows, according to the PROV-W3C model, was ensured by tools such as “prov”, “pydotplus”, and the “Graphviz” software.

The initial data acquisition occurred between January and February 2026, through a “web scraping” attempt on the Transparency Portal. However, this strategy proved unfeasible due to technical barriers, such as dynamic loading via JavaScript. Given this limitation, a manual collection of data with monthly granularity was performed, covering the period from 2020 to 2024, resulting in a dataset in CSV format.

Subsequently, the consumption of the Transparency Portal API was adopted to obtain data in JSON format, seeking greater interoperability and automation. The manually collected database was maintained concurrently with the API database, for the purpose of conducting a methodological comparison and validating the time series. This approach allowed for comparing the consistency and integrity of the data obtained by different methods.

The analysis was enriched by the survey of five Management Reports from each institution, covering the years 2020 to 2024. These documents were downloaded from the official portals of the selected universities. Manual processing of these PDF files allowed for the conversion of fragmented content into objective indicators, facilitating triangulation with quantitative data and verification of the correspondence between statistical trends and the described administrative realities.

For the 2025 budget projection, the SARIMA (Seasonal Autoregressive Integrated Moving Average) model was used, selected for its robustness in handling financial data that exhibit annual seasonality and nominal growth trends. The model was configured to capture the autocorrelations of monthly expenses, considering the settlement peaks typical of the fiscal year-end.

The SARIMA model training was performed with historical data from January 2020 to December 2023. The year 2024 was reserved for the model’s testing and cross-validation stage. The hyperparameter tuning (p, d, q) x (P, D, Q) was conducted using a “grid search” algorithm, aiming to minimize the Akaike Information Criterion (AIC).

For the model performance evaluation, the metrics MAE (Mean Absolute Error) and RMSE (Root Mean Square Error) were adopted. These metrics were used to quantify the magnitude of the deviations between the values predicted by the model and the actual commitments made by the universities, ensuring an objective evaluation of the accuracy of the budget projections.

3. Results and Discussion

The analysis of the budgetary dynamics of the federal universities UFRRJ and UFRJ, between 2020 and 2024, revealed a structure of high fiscal rigidity, characterized by the predominance of mandatory expenses that significantly limit the capacity for institutional investment and modernization. Nominal growth in operational expenses was observed, which, however, did not translate into a real expansion of services. This increase was mostly absorbed by the pressure of fixed costs and the need to adjust ongoing contracts, especially in the post-COVID-19 pandemic scenario, according to the literature on public governance (Reis et al., 2024).

In the case of the Federal Rural University of Rio de Janeiro (UFRRJ), the total sum of committed expenses showed a nominal growth trend, increasing from R$ 739,337,155.43 in 2020 to R$ 953,843,157.64 in 2024. This increase, although significant in absolute terms, reflects the constant inflationary pressure on costs and the adjustments of essential contracts for maintaining in-person activities. The prioritization of operational costs limited the margin for new investments in technological and physical infrastructure, compromising the institution’s long-term sustainability.

The Federal University of Rio de Janeiro (UFRJ), in turn, operated on a substantially larger budgetary scale, with a volume of resources approximately five times greater than that of UFRRJ in 2024. The amount committed by UFRJ reached BRL 4,208,289,826.32 in 2020, undergoing a nominal reduction of 3.2% in 2021, which represented a contraction of over BRL 136 million. This decrease directly impacted the capacity for new investments. In subsequent years, there was a progressive recovery, with an average nominal expansion rate of 5.1% per year between 2021 and 2024, culminating in BRL 4,728,046,132.52 committed in 2024. This growth reflects the pressure of fixed costs and ongoing contracts in a highly complex structure, which includes a vast network of hospitals and research centers.

The analysis of UFRRJ’s budgetary execution, segmented by Expense Group, revealed an extremely rigid financial structure, with a predominance of mandatory expenses. Personnel and Social Charges expenses consolidated as the main budgetary component, reaching a peak of R$ 775,391,673.51 in 2024. The Other Current Expenses group also grew from R$ 86.6 million in 2021 to R$ 176.08 million in 2024, indicating a pressure of fixed costs that surpassed inflationary adjustment. In contrast, capital investments showed marked volatility and decline, falling from R$ 9.8 million in 2022 to just R$ 2.3 million in 2024, evidencing the vulnerability of modernization resources.

At UFRJ, the budgetary structure also demonstrated high fiscal rigidity, concentrated in mandatory expenses. The central axis of the budget was Personnel and Social Charges, which reached R$ 3,704,889,213.69 in 2024, an increase of 8.2% compared to 2020. Other Current Expenses expanded by 35.3%, reaching R$ 986,521,670.17 to support campuses and hospitals, confirming the continuous pressure of operational costs. Capital investments, on the other hand, showed instability and longitudinal decline, with the committed amount reducing by approximately 32.6% in the period, from R$ 54.3 million in 2020 to R$ 36.6 million in 2024. These values remained below 1% of the total budget, signaling severe restrictions on the university’s technological modernization and physical expansion.

The disparity between the growth of current expenses and the retraction of investments suggests that both institutions operate in a permanent crisis management regime, sacrificing innovation and physical infrastructure for basic operational maintenance. This dynamic is reinforced by the analysis of data obtained via API from the Transparency Portal for UFRRJ, which showed a gradual growth in amounts paid, with a nominal expansion of 23.5% over the period, from R$ 615.5 million in 2020 to R$ 760.2 million in 2024. This continuous increase reflects the need for growing financial contributions to maintain essential activities, even in a scenario of fiscal restriction.

A relevant methodological finding was the identification of discrepancies between the values obtained through manual collection and those extracted via API. In the case of UFRRJ, nominal variations in the Committed Value showed a growth from 3.89% in 2020 to 8.39% in 2024, resulting in an approximate difference of R$ 80 million in the last fiscal year. For UFRJ, the percentage discrepancies remained relatively stable, fluctuating between 1.15% and 1.72%, with a highlight in 2023, when the deviation reached only 0.42%. Despite the percentage stability, the absolute magnitude of the discrepancies remained high due to the size of UFRJ’s budget, with a nominal difference of approximately R$ 81.5 million in 2024. However, high convergence was observed in the values related to effective cash flow (Paid Value) for both universities, corroborating the API’s effectiveness as a reliable tool for monitoring effective disbursements.

The inclusion of the Management Reports from the evaluated institutions complemented the quantitative analysis with a qualitative and strategic approach. For UFRRJ, the document highlighted that most of the budgetary allocation was directed towards operating costs, with an increase in “Other Current Expenses” from R$ 86.6 million in 2021 to R$ 176.08 million in 2024, while investments were limited to R$ 2,368,450.00. The discrepancies in 2024 were partially explained by the payment of R$ 97,729,223.59 in Outstanding Liabilities, indicating a strategy for settling historical debts. The difference between the total committed and the amount paid was justified by administrative settlement periods and the dependence on the release of federal funds, suggesting that budgetary results should be understood within the structural limitations of public funding.

The data from the UFRJ Management Report revealed a consistent budgetary trajectory, with the settlement of most committed expenses within the same fiscal year. This practice resulted in minimal discrepancies between committed and settled amounts, evidencing financial planning aligned with concrete execution capacity. The analysis of these reports is crucial for interpreting quantitative indicators through a management logic focused on short-term institutional sustainability, in accordance with the principles of transparency and governance (Reis et al., 2024).

Time series UFRRJ and UFRJ – Trends and Sustainability

The longitudinal analysis of budgetary execution through time series for UFRRJ and UFRJ, between 2020 and 2024, allowed for the identification of the trajectory of the Committed Value. Although UFRJ operates on a substantially larger budgetary scale, both universities showed a trend of resuming nominal growth in the post-pandemic period. This chronological visualization is fundamental to diagnosing whether the observed variations reflect real budget gains or if they constitute merely a pressure exerted by the increase in fixed costs and ongoing contracts after the return to in-person activities.

The time series analysis revealed a scenario of structural financial asphyxiation, characterized by the absolute predominance of mandatory expenses over investment capacity. In both institutions, fiscal sustainability is maintained by a precarious balance, where operational costs and payroll consume almost all available resources. At UFRRJ, the proportional pressure is even more evident, with a constant positive inclination in current expenses and cost volatility, suggesting management under constant cash flow risk. The most alarming data lies in the almost non-existence of monthly investments, which appear only as statistical noise.

At UFRJ, the time series is marked by stability at high levels, with seasonal peaks in personnel expenses that exert continuous pressure on the budget. Although costs show a gradual upward trend, investments suffer the most severe compression. The university’s capital trajectory, after 2020, has been reduced to residual values, transforming infrastructure modernization into a sporadic and insufficient event for the institution’s magnitude. This absence of continuous resource flow for infrastructure makes long-term planning impossible and compromises technological renewal, an essential pillar for maintaining academic excellence.

The analysis of anomalies in UFRRJ’s expenses revealed high sensitivity to specific events. A peak in investments of R$ 671 thousand in September 2022 was classified as an extraordinary investment, suggesting the occurrence of specific and non-recurring events, such as the delivery of laboratory equipment or the payment of construction measurements. The detection of outliers in the cost category pointed to a strong concentration of spending at the end of fiscal periods, with a peak of R$ 19.4 million in December 2024, reflecting an accelerated effort in financial execution. The abrupt drops in personnel expenses in January do not indicate irregularities, but rather a pattern intrinsic to the federal government’s cash flow, related to the transition between fiscal years and year-end payments.

At UFRJ, anomaly detection revealed a high concentration of liquidations in the months of fiscal year-end closing, evidencing a strategy to avoid the loss of appropriations. In November 2020, an atypical peak of R$ 16.2 million in investments was recorded, a movement that was repeated in December 2024, suggesting concentrated payment for infrastructure projects. A pressure for significant liquidation in operating expenses in December 2024, with an outlier of R$ 111.6 million, indicated an extraordinary mobilization of resources for the settlement of pending services. Anomalies in the personnel category, with significant drops in January, are a direct reflection of accounting processing and disbursement peaks in December, consolidating a structural characteristic of federal cash flow.

The cyclicity and anomalies identified in the time series, characterized by the “December effect” and the constant suppression of investments, are supported by the Theory of Budgetary Rigidity and Fiscal Incrementalism (Wildavsky, 1964). Lima and Ferreira (2024) highlight that public funding for higher education in Brazil is marked by severe instabilities and a lack of continuous resource flow. The practice of systematic compression of the Investment Group (G4) acts as an accounting “adjustment variable,” corroborating the thesis that university management operates under a model of chronic fiscal uncertainty, in which short-term financial sustainability compromises long-term physical and technological assets.

The irregular component of UFRRJ demonstrated moderate volatility, with significant residual peaks reaching R$ 9.7 million outside the expected pattern. These “noises” suggest dependence on extraordinary events, such as supplementary transfers or delays in federal transfers, destabilizing linear planning and making financial management more reactive to random external elements than truly strategic. At UFRJ, the magnitude of the residuals is higher, reflecting the complexity of managing a billion-dollar structure subject to random variations of up to R$ 24.5 million. The random dispersion of these values, without a clear pattern, highlights a fiscal uncertainty environment in which non-seasonal events, such as sharp variations in university hospital revenue or unexpected judicial executions, challenge budgetary predictability.

Budget Planning and Projection

The budget forecast for UFRRJ points to a challenging year in 2025, with a projected resource requirement of approximately R$ 845.6 million. The slope of the trend curve at UFRRJ is proportionally steeper than at UFRJ, reflecting the direct impact of contract inflation and the vegetative growth of payroll on a smaller-scale budget. The statistical model anticipates that the university’s characteristic volatility will persist, requiring special attention to the transition months between quarters, where residual deviations tend to be more frequent and can impact the institution’s immediate payment capacity.

The prospective analysis shows that UFRRJ’s fiscal sustainability depends on a recomposition that considers the real increase in basic maintenance costs, which have doubled in recent years. Without this adjustment, the projection indicates that the university may face difficulties in maintaining the current level of services, with infrastructure investments remaining at near-zero levels. Therefore, the projection serves as a warning tool for managers, signaling that the strategy of using investment as an adjustment variable has reached its technical limit, making it imperative to seek supplementary sources of funding to ensure future academic excellence.

Statistical modeling for UFRJ projects a scenario of continuity in pressure on the budgetary ceiling, with an estimate of total disbursement for the fiscal year 2025 on the order of R$ 4.39 billion. The projection captures the maintenance of seasonal personnel payment cycles and the trend of increase in operational cost expenses, indicating that the institution will operate at a higher financial level than in previous years. The confidence interval of the forecast suggests that, although possible variations may occur due to administrative unforeseen events, the baseline of mandatory expenses is consolidated at levels that require an uninterrupted flow of federal transfers to avoid the shortage of essential services.

This upward trend reinforces the need for risk management that anticipates the settlement peaks identified for the end of the year. The forecast indicates that the margin for infrastructure investments will remain extremely limited if there is no change in the total budgetary allocation, as projected current expenses tend to absorb almost all of the planned resources. Planning for 2025 must prioritize spending efficiency for operating costs to try to make room for corrective maintenance of the facilities, at the risk of deepening the physical obsolescence already observed in the historical period.

The methodological architecture developed in this work transcends academic analysis, offering practical tools applicable to planning secretariats, administration pro-rectorates, and internal and external control bodies. One of the research products is the Automated Monitoring and Early Warnings, which integrates predictive models (SARIMA) and provenance traceability, allowing university managers to simulate budget settlement scenarios. This tool identifies in advance the risk of operational contract shortages or investment stagnation. The Python data processing routine, based on data consumed from the Transparency Portal API, can continuously feed an interactive dashboard, translating the SARIMA model’s projections into simplified indicators and managerial alerts.

Another contribution is the Standardization of Active Transparency. By highlighting and mitigating the discrepancies between raw data from web portals and API data, the research offers a protocol for federal universities to structure their open data portals according to FAIR principles (Findable, Accessible, Interoperable, Reusable), guaranteeing auditable and real-time information to society and oversight bodies (Wilkinson et al., 2016). It is important to note that the projections for 2025 were based on the historical behavior of time series from 2020 to 2024, as the consolidated official data for 2025 had not yet been published. This absence of immediate past data reinforces the relevance of using predictive models in Data Science to assist university strategic planning in the face of budgetary uncertainties.

In summary, the results demonstrate that UFRJ and UFRRJ face pronounced fiscal rigidity, with mandatory expenses consuming most of the budget and severely limiting investments in infrastructure and technology. Time series analysis and anomaly detection revealed a reactive management to specific events and a dependence on irregular cash flows, evidencing a model of chronic fiscal uncertainty. Projections for 2025 indicate the continuation of this pressure, with sustainability depending on new financing policies and robust data governance, which ensures the integrity and traceability of fiscal decisions, according to FAIR principles and the PROV-W3C model.

4. Conclusion

This study investigated the transparency and budgetary dynamics of the federal universities UFRJ and UFRRJ between the years 2020 and 2024, based on FAIR principles and the PROV-W3C provenance model to ensure the integrity of records. High fiscal rigidity was observed in both institutions, characterized by the predominance of mandatory expenses for personnel and social charges, which peaked in 2024. It was noted that the nominal growth of current expenses in the post-pandemic period did not represent a real expansion of services, but rather the pressure of fixed costs and ongoing contracts. In contrast, capital investments suffered severe contractions, evidencing a trajectory of physical and technological obsolescence. The methodological confrontation between manual collection and the API identified significant nominal discrepancies in the committed value, although there was high convergence in effective cash flows. It was concluded that the universities operated in a permanent crisis management regime, sacrificing infrastructure modernization for immediate operational survival. Time series analysis revealed a scenario of structural financial asphyxiation, with operational costs and payroll consuming almost all available resources, and the suppression of investments as a constant.

The methodological architecture developed offers significant practical contributions, such as Automated Monitoring and Early Alerts, which integrates SARIMA predictive models and provenance traceability to simulate budget settlement scenarios and identify risks of stockouts or investment bottlenecks, with the potential to feed an interactive dashboard. Additionally, the Standardization of Active Transparency proposes a protocol for universities to structure their open data portals according to FAIR principles, ensuring auditable and real-time information. As a limitation, the projections for 2025 were based on the historical behavior of time series from 2020 to 2024, given the absence of consolidated official data for the 2025 fiscal year until the closing of this work. It is recommended that future studies incorporate the 2025 data as soon as they are available, allowing for a comparison between projections and actual execution to improve the tuning of the prediction models’ hyperparameters. The financial sustainability of federal higher education institutions depends on new funding policies and robust data governance that ensures the integrity and traceability of fiscal decisions.

Bibliographic References

Lima, C.P.L.; Ferreira, M.A.S. 2024. Financiamento público estudantil no ensino superior: um olhar para a assistência estudantil no Brasil e para o serviço de apoio social em Portugal (2015-2022). Revista Brasileira de Educação 31: e310012.

Reis, E.O.; Almeida, R.A.; Melo, D.C.; Pereira, G.A. 2024. O princípio da transparência na governança pública no Brasil: uma revisão bibliométrica. Revista Facthus de Administração e Gestão 7(1): 106-131.

Wildavsky, A. The politics of the budgetary process. Boston: Little, Brown and Company, 1964.

Wilkinson, M.D.; Dumontier, M.; Aalbersberg, I.J.; Appleton, G.; Axton, M.; Baak, A.; Blomberg, N.; Boiten, J.W.; Santos, L.B.S.; Bourne, P.E.; Bouwman, J.; Brookes, A.J.; Clark, T.; Crosas, M.; Dillo, I.; Dumon, O.; Edmunds, S.; Evelo, C.T.; Finkers, R.; Gonzalez-Beltran, A.; Gray, A.J.G.; Groth, P.; Goble, C.; Grethe, J.S.; Heringa, J.; ‘t Hoen, P.A.C.; Hooft, R.; Kuhn, T.; Kok, R.; Kok, J.; Lusher, S.J.; Martone, M.E.; Mons, A.; Packer, A.L.; Persson, B.; Rocca-Serra, P.; Roos, M.; van Schaik, R.; Sansone, S.A.; Schultes, E.; Sengstag, T.; Slater, T.; Strawn, G.; Swertz, M.A.; Thompson, M.; van der Lei, J.; van Mulligen, E.; Velterop, J.; Waagmeester, A.; Wittenburg, P.; Wolstencroft, K.; Zhao, J.; Mons, B. 2016. The FAIR Guiding Principles for scientific data management and stewardship. Scientific Data 3(160018): 1-9.

World Wide Web Consortium [W3C]. 2013. PROV Overview: An overview of the PROV family of documents. Disponível em: https://www.w3.org/TR/prov-overview/. Acesso em: 12 dez. 2025.

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