Students and parents had to make challenging decisions regarding education during the COVID-19 pandemic. School districts across the US implemented varied school operation plans and the federal government did not create a centralized source for operational statuses at the national level. Data at the state level differs greatly in availability, verbiage, and format. This project aggregates and tracks district-level school operational statuses state-by-state, as well as school-related COVID-19 cases. It accounts for 6,800 out of 10,000 US school districts across 33 states. With this data, K-12 parents and children have a tool to make informed decisions regarding education and health.

Financial institutions have long been looking for quantitative methods to analyze tens of thousands of jargon-filled documents generated in the highly dynamic market. Thus, MarketOwl is the solution that can collect and assimilate these text data into measurable indicators. Through analyzing SEC filings and earnings call transcripts with its NLP engine, MarketOwl provides insights such as sentiment and credibility trends and transforms them into curated dashboards. MarketOwl empowers portfolio managers to make faster and better decisions with integrated and trustworthy company information and make more profits for their company and clients.

Our team worked with the Microsoft Azure Machine Learning product team to improve the retention of deeply engaged customers on the AML platform. The project involved both qualitative and quantitative analysis of customer usage patterns, including customer interviews, telemetry data analysis, hypothesis building and testing. The result of our analysis serves as a good foundation to understand the needs and pain points of the customers and to come up with strategic ways of moving forward. It has the potential to help improve the user experience of the AML platform and allow AML to gain more competitive advantages in the market.

With more than 1 billion people with disabilities worldwide, we believe that accessibility is essential and should not be left as an afterthought. Data visualizations are one area where accessibility lags behind. This is especially important for Power BI, a growing platform with over 5 million subscribers. Power BI creators have difficulties building accessible reports due to a lack of accessibility tools & standard guidelines on how to make visualizations accessible. Thus, we developed PowBility: a chrome extension that enables Power BI report creators to design accessible visualizations. PowBility aims to enable everyone to utilize visualizations to their full potential.

Epic’s Unreal Engine is complex with a lot of documentation and courses. The content creators face difficulty finding the information they are looking for and thus, require dynamically recommended related documentation/courses based on the Unreal Online Learning (UOL) search. We built a predictive analytics tool to recommend similar documentation and courses based on the UOL journey map of the user, working with terms from the taxonomies and knowledge graph. Our project helps users from different backgrounds to identify documentation easily, enabling equitable access to the tool. It also provides a personalized search experience for the user.

A significant challenge for people with disabilities is the lack of awareness around accessible employers, programs, and opportunities within the community. Additionally, the information they seek on inclusive hiring practices, benefits, accommodations, and company culture is scattered across multiple sources. Our application uses data mining techniques on company blogs, Twitter, Glassdoor, YouTube, etc., to aggregate such information, and, in doing so, empowers the underserved disabled community to make more informed job-search decisions. The application also aims to help employers identify gaps in the information they provide, thus bringing them closer to diverse candidates' needs.

The security industry is rapidly developing and changing, creating a knowledge gap and skills shortage due to the challenging learning curve. Analysts and students struggle with the overwhelming amount of data that typical network analysis tools display, which leads to a time-consuming investigation process. The SecViz tool allows SecOps analysts to read and dynamically visualize packet logs, leading to faster identification, investigation, and mitigation of KPIs. In addition, the tool helps expedite the speed of identifying security gaps that may have been missed or taken a significant amount of time to come across.

The current quarterly access review process at Smartsheet comes with a myriad of challenges. Smartsheet’s access reviews are time-consuming and draw upon a lot of resources. There’s a large company-wide communication effort to track changes in employee structure. Our Identity and Access Management Dashboard presents the Smartsheet Compliance team with employee change metrics tracked with custom automation, a report containing highlighted daily employee changes, and visualizations used to present a historical timeline of changes in employee structure. With our dashboard, Smartsheet’s Compliance team will be able to save time and resources and have key actionable insights to employee changes.

The Sno-King Amateur Hockey Association collects and stores tens of thousands of registration and donation data entries across seven different platforms each year. With their data storage not being centralized, accessing their data has become increasingly difficult as the organization has grown and more data continues to flow in. For our Capstone project, we were tasked with creating and filling a centralized database, and then making analyses based on that data. From the work we have done, our sponsor will be able to make their own analyses to make better marketing decisions for its adult and youth hockey programs.