Much research focuses on attrition out of career paths in the fields of science, technology, engineering, and mathematics (STEM), but many students change from one career path to another within STEM (e.g., changing from a PhD researcher to a lab technician). This study examined what motivated college students to change career plans within STEM, with particular attention to which motivational concern(s) (as defined by situated expectancy-value theory) students reported considering when reflecting on decision making, both individually and in combination. As part of a larger study, we identified 561 college students who changed career plans within STEM at a large 4-year research university. On average, students reported attractive aspects related to attainment/utility or intrinsic value of their new career paths to be the most salient influences on their decisions to change plans within STEM. However, virtually all students referenced multiple motivational concerns, related to both their new and old careers, as shaping their decisions in some way. Latent class analyses suggested three distinct patterns of motivational concerns that students reported considering in combination: some students (53.3%) reported a predominant focus on value-related concerns, some (36.2%) reported considering multifaceted motivational concerns, and some (10.5%) reported a predominant focus on competence-related concerns. There were few differences as a function of gender, race, year in school, and field of study, but students with lower college grade point averages more often reported considering competence-related concerns during decision making. Results point to the importance of examining the multiple motivational concerns that college students negotiate simultaneously (including motivation for multiple different career options) during career decision making.

Background Research and policy often focus on reducing attrition from educational trajectories leading to careers in science, technology, engineering, and mathematics (STEM), but many students change career plans within STEM. This study examined how changing career plans within STEM fields was associated with psychological indicators of career readiness. We conducted a large online survey of undergraduate students ( N = 1,727) across 42 courses covering every major STEM discipline at a large U.S. research-intensive public university. Students reported about their career plans, whether plans had changed, motivation for those career plans, and satisfaction with and certainty of persisting with those plans. A trained team of coders classified whether students reported having STEM career plans at the time of the survey and at the beginning of college. Results Students who said they had changed career plans within STEM fields during college also reported lower motivation for their new career plans, satisfaction with those plans, and certainty of persisting in them, compared to students who retained consistent STEM career plans. With few exceptions, these associations held across students’ gender, race, year in school, and STEM field of study. Within-STEM career plan changes were very common, reported by 55% of fourth-year STEM students. Women reported changing career plans within STEM fields more often than men. Implications Results suggest that changing career plans within STEM is an important phenomenon to consider in preparing a qualified and diverse STEM workforce. Students who change career plans within STEM fields may need additional supports for their career motivation and satisfaction compared to students who do not change plans.

Over the past decade, active learning pedagogical approaches have increased in popularity among multiple STEM disciplines, including Computer Science (CS). The purpose of this experience report is to reflect on the efforts taken to incorporate active learning into some of the undergraduate CS courses at the University of Georgia. This report describes the implementation details of two courses, CS2 and Discrete Mathematics, as taught during the Spring semester in 2019 with a total initial enrollment count of 551, and it summarizes the results of a common set of anonymous exit survey responses collected by the instructors on a subset of that population. In these courses, active learning was implemented using a flipped classroom model with undergraduate peer learning assistants. The surveyed population consisted of 163 CS2 students and 125 Discrete Mathematics students. The responses suggest that students either understand or believe that the active learning activities they participated in helped with their learning, especially regarding confidence and question formation. However, 67.36% of responses reported a high degree of state anxiety when receiving new assignments. Additionally, while 89.58% of responses expressed a desire to take a class with peer learning assistants in the future, only 40.63% indicated a preference for active learning when compared to traditional lecture. The course instructors discuss how they will use the insights ascertained via the exit surveys as evidence when deciding changes for future iterations of the courses.

In recent years, active learning as a pedagogical approach has increased in popularity in Computer Science (CS) education and other Science, Technology, Engineering, and Math (STEM) disciplines; the number of search results for "active learning" in the ACM Digital Library has roughly doubled from 2014 to 2019 alone. A recent study finds that when active learning is introduced in some STEM classrooms, students feel like their performance drops when empirical evidence shows that it actually increases [3]. Some of our previous work for Spring 2019 [2] anecdotally confirms those findings and prompted us to make changes for Fall 2019. The purpose of this poster is to report and reflect on those changes. The results suggest an improvement in student sentiment regarding active learning over those semesters. In this work, the term active learning refers to the approach described in Collins and O'Brien [1]; and peer-based active learning refers an active learning environment where undergraduate students called "Peer Learning Assistants" (PLAs) are utilized to help answer student questions during activities [4]. This poster describes the implementation details of two courses, CS2 and Discrete Mathematics, as taught during the Spring and Fall semesters in 2019 with a total initial enrollment count of 551 in the Spring and 490 in the Fall. It summarizes the results of a common set of anonymous exit survey responses collected by the instructors on a subset of that population, and it discusses the specific changes the course instructors decided to make based on the insights ascertained from the first semester survey. In these courses, active learning was implemented using a flipped classroom model with undergraduate PLAs. The surveyed population consisted of 163 CS2 students and 125 Discrete Mathematics students in Spring 2019 and 190 CS2 students and 140 Discrete Mathematics students in Fall 2019. The subset of questions presented in this poster are the following: (Q1) Given the option, would you rather work in small student groups on a problem in lecture (active learning) or watch the instructor solve the problem (traditional lecture)? (Q2) I felt comfortable asking questions to the professor. (Q3) After interacting with the Undergrad PLA(s), I feel like I ask better questions. Responses are summarized in Table 1. The poster describes how the structure and presentation of course materials scaffold learning outcomes and allow students to build up towards higher stakes assignments. It also discusses some of the specific changes that the instructors made to both courses between Spring 2019 and Fall 2019, which are outlined below: i) incorporating strategies from the Transparency in Learning and Teaching (TILT) project; ii) using innovative, evidence-based instructional practices in accordance with our university's initiative for active learning; iii) further utilizing an undergraduate assistant experiential learning program where PLAs are hired and paid by the department and incorporating PLAs into all affected course sections; and iv) specifically focusing on morale early on by discussing past experiences and research results related to peer-based active learning with the students. These changes combined with survey results suggest that there are concrete steps that can be taken to improve student sentiment of active learning.

Forecasting is the art of taking available information of the past and attempting to make the best educated guesses of the ever unforeseen future. From the historical data, patterns can be observed and forecasting models have been developed to capture such patterns. This work focuses on forecasting traffic flow in major urban areas and freeways in the state of Georgia using large amounts of data collected from traffic sensors. Much of the existing work on traffic flow forecasting focuses on the immediate short terms. In addition to that, this work studies the forecasting powers of various models, including seasonal ARIMA, exponential smoothing and neural networks, for relatively long terms. A second experiment that incorporates precipitation data into forecasting models to better predict traffic flow in rainy weather is also conducted. Dynamic regression models and neural networks are used in this experiment. In both experiments, neural networks outperformed the others overall.

Open Science Big Data is emerging as an important area of research and software development. Although there are several high quality frameworks for Big Data, additional capabilities are needed for Open Science Big Data. These include data provenance, citable reusable data, data sources providing links to research literature, relationships to other data and theories, transparent analysis/reproducibility, data privacy, new optimizations/advanced algorithms, data curation, data storage and transfer. An important part of science is explanation of results, ideally leading to theory formation. In this paper, we examine means for supporting the use of theory in big data analytics as well as using big data to assist in theory formation. One approach is to fit data in a way that is compatible with some theory, existing or new. Functional Data Analysis allows precise fitting of data as well as penalties for lack of smoothness or even departure from theoretical expectations. This paper discusses principal differential analysis and related techniques for fitting data where, for example, a time-based process is governed by an ordinary differential equation. Automation in theory formation is also considered. Case studies in the fields of computational economics and finance are considered.

Predictive analytics in the big data era is taking on an ever increasingly important role. Issues related to choice on modeling technique, estimation procedure (or algorithm) and efficient execution can present significant challenges. For example, selection of appropriate and optimal models for big data analytics often requires careful investigation and considerable expertise which might not always be readily available. In this paper, we propose to use semantic technology to assist data analysts and data scientists in selecting appropriate modeling techniques and building specific models as well as the rationale for the techniques and models selected. To formally describe the modeling techniques, models and results, we developed the Analytics Ontology that supports inferencing for semi-automated model selection. The SCALATION framework, which currently supports over thirty modeling techniques for predictive big data analytics is used as a testbed for evaluating the use of semantic technology.

Predictive analytics in the big data era is taking on an ever increasingly important role. Issues related to choice on modeling technique, estimation procedure/algorithm and efficient execution can present significant challenges. For example, selection of appropriate/optimal models for big data analytics often requires careful investigation and considerable expertise which might not always be readily available. In this paper, we propose to use semantic technology to assist data analysts/data scientists in selecting appropriate modeling techniques and building specific models as well as the rationale for the techniques and models selected. To formally describe the modeling techniques, models and results, we developed the Analytics Ontology that supports inferencing for semi-automated model selection. The ScalaTion framework, which currently supports over thirty modeling techniques for predictive big data analytics is used as a test bed for evaluating the use of semantic technology.

Predictive analytics and simulation modeling are two complementary disciplines that will increasingly be used together in the future. They share in common a focus on predicting how systems, existing or proposed, will function. The predictions may be values of quantifiable metrics or classification of outcomes. Both require collection of data to increase their validity and accuracy. The coming era of big data will be a boon to both and will accelerate the need to use them in conjunction. This paper discusses ways in which the two disciplines have been used together as well as how they can be viewed as belonging to the same modeling continuum. Various modeling techniques from both disciplines are reviewed using a common notation. Finally, examples are given to illustrate these notions.

The explosion of available data along with the need to integrate and utilize that data has led to a pressing interest in data integration techniques. In terms of Semantic Web technologies, Ontology Alignment is a key step in the process of integrating heterogeneous knowledge bases. In this paper, we present the Edge Confidence technique, a modification and improvement over the popular Similarity Flooding technique for Ontology Alignment.

In order to optimally schedule energy, consumers need to understand their energy utility function. The traditional utility function must be modified when consumers acquire production capabilities and become prosumers. This paper presents a formal economic model for prosumers, appropriately integrating distributed generation, storage and demand response capabilities. We identify the limitations of existing prosumer-like models. Finally, we demonstrate how the proposed prosumer model is applied to energy evaluation to distributed prosumer decisions regarding environmental objectives.

The explosion of available data along with the need to integrate and utilize that data has led to a pressing interest in data integration techniques. In terms of Semantic Web technologies, Ontology Alignment is a key step in the process of integrating heterogeneous knowledge bases. In this paper, we present the Edge Confidence technique, a modification and improvement over the popular Similarity Flooding technique for Ontology Alignment.

The wide-scale development of ontologies in the bioinformatics domain facilitates their use in the creation of scientific workflows. To speed up the design of workflows, a Service Suggestion Engine is interfaced to the Galaxy Tool Integration and Workflow Platform. This enables users to ask for suggestions (e.g., what operation should go next) while designing workflows with the Galaxy user interface. The Service Suggest Engine utilizes semantic annotations to suggest appropriate Web service operations to plug into the workflow under design. The enriched Ontology for Biomedical Investigation (OBI) is used as a target for the annotations. The effectiveness of the suggestions provided is evaluated against a consensus of domain experts.

Determining the value of distributed energy storage for grid operation and control requires modeling and simulation of optimal energy scheduling. This paper evaluates Stochastic Dynamic Programming (SDP) and Stochastic Game Theory (SGT) methods for scheduling energy usage in a system with storage, and quantitatively evaluates these scheduling methods for energy storage in commercial buildings. With varying probabilities of load, an optimal energy storage strategy is derived that provides a way to determine the value of load forecasting on a daily basis and evaluates this strategy in the face of energy use incentives. Theoretical results are matched with modeling and simulation.

Just as the other informatics-related domains (e.g., Bioinformatics) have discovered in recent years, the ever-growing domain of Energy Informatics (EI) can benefit from the use of ontologies, formalized, domain-specific taxonomies or vocabularies that are shared by a community of users. In this paper, an overview of the Ontology for Energy Investigations (OEI), an ontology that extends a subset of the well-conceived and heavily-researched Ontology for Biomedical Investigations (OBI), is provided as well as a motivating example demonstrating how the use of a formal ontology for the EI domain can facilitate correct and consistent knowledge sharing and the multi-level analysis of its data and scientific investigations.

As recent programming languages provide improved conciseness and flexibility of syntax, the development of embedded or internal Domain-Specific Languages has increased. The field of Modeling and Simulation has had a long history of innovation in programming languages (e.g. Simula-67, GPSS). Much effort has gone into the development of Simulation Programming Languages. The ScalaTion project is working to develop an embedded or internal Domain-Specific Language for Modeling and Simulation which could streamline language innovation in this domain. One of its goals is to make the code concise, readable, and in a form familiar to experts in the domain. In some cases the code looks very similar to textbook formulas. To enhance readability by domain experts, a version of ScalaTion is provided that heavily utilizes Unicode. This paper discusses the development of the ScalaTion DSL and the underlying features of Scala that make this possible. It then provides an overview of ScalaTion highlighting some uses of Unicode. Statistical analysis capabilities needed for Modeling and Simulation are presented in some detail. The notation developed is clear and concise which should lead to improved usability and extendibility.