Journal of Data Science

This paper presents an in-depth analysis of patterns and trends in the open-source software (OSS) contributions by the U.S. federal government agencies. OSS is a unique category of computer software notable for its publicly accessible source code and the rights it provides for modification and distribution for any purpose. Prompted by the Federal Source Code Policy (USCIO, 2016), Code.gov was established as a platform to facilitate the sharing of custom-developed software across various federal government agencies. This study leverages data from Code.gov, which catalogs OSS projects developed and shared by government agencies, and enhances this data with detailed development and contributor information from GitHub. By adopting a cost estimation methodology that is consistent with the U.S. national accounting framework for software investment proposed in Korkmaz et al. (2024), this research provides annual estimates of investment in OSS by government agencies for the 2009–2021 period. The findings indicate a significant investment by the federal government in OSS, with the 2021 investment estimated at around $407 million. This study not only sheds light on the government’s role in fostering OSS development but also offers a valuable framework for assessing the scope and value of OSS initiatives within the public sector.

Text clustering can streamline many labor-intensive tasks, but it creates a new challenge: efficiently labeling and interpreting the clusters. Generative large language models (LLMs) are a promising option to automate the process of naming text clusters, which could significantly streamline workflows, especially in domains with large datasets and esoteric language. In this study, we assessed the ability of GPT-3.5-turbo to generate names for clusters of texts and compared these to human-generated text cluster names. We clustered two benchmark datasets, each from a specialized domain: research abstracts and clinical patient notes. We generated names for each cluster using four prompting strategies (different ways of including information about the cluster in the prompt used to get LLM responses). For both datasets, the best prompting strategy beat the manual approach across all quality domains. However, name quality varied by prompting strategy and dataset. We conclude that practitioners should consider trying automated cluster naming to avoid bottlenecks or when the scale of the effort is enough to take advantage of the cost savings offered by automation, as detailed in our supplemental blueprint for using LLM cluster naming. However, to get the best performance, it is vital to test a variety of prompting strategies and perform a small test to identify which one performs best on each project’s unique data.

The coronavirus disease 2019 (COVID-19) pandemic presented unique challenges to the U.S. healthcare system, particularly for nonprofit U.S. hospitals that are obligated to provide community benefits in exchange for federal tax exemptions. We sought to examine how hospitals initiated, modified, or disbanded community benefits programming in response to the COVID-19 pandemic. We used the free-response text in Part IV of Internal Revenue Service (IRS) Form 990 Schedule H (F990H) to assess health equity and disparities. We combined traditional key term frequency and Hierarchical Density-Based Spatial Clustering of Applications with Noise (HDBSCAN) clustering approaches with a novel Generative Pre-trained Transformer (GPT) 3.5 summarization approach. Our research reveals shifts in community benefits programming. We observed an increase in COVID-related terms starting in the 2019 tax year, indicating a pivot in community focus and efforts toward pandemic-related activities such as telehealth services and COVID-19 testing and prevention. The clustering analysis identified themes related to COVID-19 and community benefits. Generative Artificial Intelligence (GenAI) summarization with GPT3.5 contextualized these changes, revealing examples of healthcare system adaptations and program cancellations. However, GPT3.5 also encountered some accuracy and validation challenges. This multifaceted text analysis underscores the adaptability of hospitals in maintaining community health support during crises and suggests the potential of advanced AI tools in evaluating large-scale qualitative data for policy and public health research.

The North American Product Classification System (NAPCS) was first introduced in the 2017 Economic Census and provides greater detail on the range of products and services offered by businesses than what was previously available with just an industry code. In the 2022 Economic Census, NAPCS consisted of 7,234 codes and respondents often found that they were unable to identify correct NAPCS codes for their business, leaving instead written descriptions of their products and services. Over one million of these needed to be reviewed by Census analysts in the 2017 Economic Census. The Smart Instrument NAPCS Classification Tool (SINCT) offers respondents a low latency search engine to find appropriate NAPCS codes based on a written description of their products and services. SINCT uses a neural network document embedding model (doc2vec) to embed respondent searches in a numerical space and then identifies NAPCS codes that are close to the search text. This paper shows one way in which machine learning can improve the survey respondent experience and reduce the amount of expensive manual processing that is necessary after data collection. We also show how relatively simple tools can achieve an estimated 72% top-ten accuracy with thousands of possible classes, limited training data, and strict latency requirements.

The US Department of Agriculture (USDA) National Agricultural Statistics Service (NASS) has begun a modernization effort to supplement survey data with non-survey data to improve estimation of agricultural quantities. As part of this effort, NASS has begun georeferencing farms on its list frame by linking geospatial data on agricultural fields with farm records on the list frame. Although many farms can be linked to geospatial data acquired by the Farm Service Agency (FSA), this linkage is not possible for farmers who do not participate in FSA programs, which may include members of some underrepresented groups in US agriculture. Thus, NASS has developed a georeferencing process for non-FSA farms, combining automatic and manual field identification, county assessor parcel data, record linkage, and classification surveys. This process serves the dual purpose of linking farms already on the list frame to geospatial data sources and identifying new farms to add to NASS’s list frame. This report evaluates the output of the non-FSA georeferencing process for 11 states, with a focus on farms added to the list frame via georeferencing. Substantial percentages (>25% for each category) of the new farms added via georeferencing were urban or suburban farms, were small, had livestock, or were in counties with Amish settlements. The georeferencing process shows promise adding farms from these groups that have historically been less well covered in NASS surveys.

The presence of outliers in a dataset can substantially bias the results of statistical analyses. In general, micro edits are often performed manually on all records to correct for outliers. A set of constraints and decision rules is used to simplify the editing process. However, agricultural data collected through repeated surveys are characterized by complex relationships that make revision and vetting challenging. Therefore, maintaining high data-quality standards is not sustainable in short timeframes. The United States Department of Agriculture’s (USDA’s) National Agricultural Statistics Service (NASS) has partially automated its editing process to improve the accuracy of final estimates. NASS has investigated several methods to modernize its anomaly detection system because simple decision rules may not detect anomalies that break linear relationships. In this article, a computationally efficient method that identifies format-inconsistent, historical, tail, and relational anomalies at the data-entry level is introduced. Four separate scores (i.e., one for each anomaly type) are computed for all nonmissing values in a dataset. A distribution-free method motivated by the Bienaymé-Chebyshev’s inequality is used for scoring the data entries. Fuzzy logic is then considered for combining four individual scores into one final score to determine the outliers. The performance of the proposed approach is illustrated with an application to NASS survey data.

Missing data is a common occurrence in various fields, spanning social science, education, economics, and biomedical research. Disregarding missing data in statistical analyses can introduce bias to study outcomes. To mitigate this issue, imputation methods have proven effective in reducing nonresponse bias and generating complete datasets for subsequent analysis of secondary data. The efficacy of imputation methods hinges on the assumptions of the underlying imputation model. While machine learning techniques such as regression trees, random forest, XGBoost, and deep learning have demonstrated robustness against model misspecification, their optimal performance may necessitate fine-tuning under specific conditions. Moreover, imputed values generated by these methods can sometimes deviate unnaturally, falling outside the normal range. To address these challenges, we propose a novel Predictive Mean Matching imputation (PMM) procedure that leverages popular machine learning-based methods. PMM strikes a balance between robustness and the generation of appropriate imputed values. In this paper, we present our innovative PMM approach and conduct a comparative performance analysis through Monte Carlo simulation studies, assessing its effectiveness against other established methods.