Ally Financial: Digital Avatar Assistant
Amazon: Amazon Web Services: AWSome Availability Zones
Anthropocene Institute: Air Pollution Health Outcomes Forecasting Tool
Anthropocene Institute: Electricity Grid Planning Tool
Atomic Object: Stroodle: Learning Management System
Auto-Owners Insurance: Yard Wars: Weathering the Storm
Bosch: Hardware in the Loop (HIL) Vehicle Simulator
Delta Dental of Michigan, Ohio, and Indiana: Smart Benefit Plan Recommender Engine
Delta Dental of Michigan, Ohio, and Indiana: Microsoft Excel Data Extractor / Modeler
The Dow Chemical Company: Virtual Computer Service Enhancements
Evolutio: ERP Kids: Wildlife Conservation
Ford Motor Company: Crowd Sourced EV Emergency Recharge
General Motors: Enhanced MISP User Interface
Herman Miller: Live Platform CAD Ingestion
Lockheed Martin Space: SmartSat Satellite App Store
Malleable Minds: Review Aggregator for Educational Programs
Meijer: mHealthy: Healthy Eating Application
Microsoft: Feedback Prompt for Ratings in Google Play Store
Mozilla Corporation: Improve High Contrast Mode for Firefox
MSU Federal Credit Union: Spaving: Giving based on Spending Habits
PwC: Collaboration Bot for Microsoft Teams
The Rocket Companies: ROCKY: Team Challenge Application
Stellantis: Interactive Digital Assistant
TechSmith: Snagit Template Creator
United Airlines: Gate Hazard Geo-Mapping
United Airlines: QA Audit Center
Urban Science: Independent Repair Facility (IRF) Insights
Vectorform: Smart Auto-Time Logging
Volkswagen Group of America: Car-Net DriveView Social Competition App
Whirlpool Corporation: AI Recipe Converter
Ally Financial: Digital Avatar Assistant
Ally Financial is a financial services company based in Detroit, Michigan, operating as one of the largest car finance companies in the United States. Ally has amassed an immense customer base, financing cars for over 4 million people and having 2 million depositors. Ally also offers online banking and online trading, bolstering the services they provide for their customers.
Ally is a strictly digital company, offering no physical locations for customers. Because of this, Ally has been innovating the online service field since their founding. As artificial intelligence (AI) has advanced, Ally has been exploring inventive methods of digital customer service rooted in AI.
Our Digital Avatar Assistant is a cutting-edge AI assistant that provides Ally customers with real-time communication relating to their accounts, as well as custom-tailored financial advice.
Customers interact with our assistant through a chatbot interface embedded in Ally’s website. Users ask questions by typing or speaking with the assistant about a wide array of topics, including account information, budgeting, spending analysis, etc.
To provide a lifelike experience, our assistant reacts with animated movement and facial expressions depending on context, mimicking a conversation with a human.
We use machine learning, natural language processing, and AI to analyze customer spending and budgeting habits to provide user-specific financial advice. Our Digital Avatar Assistant keeps track of user spending and provides reminders and warnings if the user is in danger of not meeting their spending goals.
The Digital Avatar Assistant is developed using Rasa. Our application uses Amazon EC2 for machine learning, Amazon S3 for model storage, DynamoDB for conversation storage, and a combination of Amazon Transcribe and Amazon Polly for conversational functionality. The user interface is built with React.
Amazon: Amazon Web Services: AWSome Availability Zones
Founded in Bellevue, Washington in 1994, Amazon is a Fortune 500 company that provides a variety of services to customers as the world’s largest online retailer and cloud services provider.
Customers using Amazon’s cloud platform, Amazon Web Services (AWS), can choose to break their application up into many parts, each hosted in a different location (called an Availability Zone, or AZ). Such redundancy helps prevent service outages for customers.
Distributed applications are spread across multiple servers, which need to communicate with each other for the application to function. This communication can take a significant amount of time, and minimizing the delay can lead to a better user experience. Choosing which AZs to use to minimize this delay traditionally requires extensive manual testing.
Our AWSome Availability Zones web application continuously and automatically measures the delay between Availability Zones, allowing Amazon Web Services customers to easily choose the fastest Availability Zones for their application, saving them time and money.
AWSome Availability Zones provides customers with an easy- to-understand visualization of the delay between Availability Zones using an interactive map with a familiar look and feel.
Experienced Amazon Web Services customers can opt to use our AWSome Availability Zones system to explore more detailed views of the network latency data, allowing them to answer specific questions they have, quickly and seamlessly.
Our software’s front end is built using Angular, and its back end uses Amazon Web Services Elastic Compute Cloud (EC2) instances to measure network latency between Availability Zones, which it stores in DynamoDB.
Anthropocene Institute: Air Pollution Health Outcomes Forecasting Tool
The Anthropocene Institute is an organization that partners with researchers, governments, experts and investors to address one of humanity’s most pressing concerns, climate change. The organization provides support to projects related to clean energy, anti- pollution efforts and climate innovation and brings down any political or financial barriers they may experience.
The Anthropocene Institute has turned its attention towards air pollution in hopes of researching the effect that air quality has on premature deaths and health complications, such as increased asthma, infant mortality and lung cancer.
Our Air Pollution Health Outcomes Forecasting Tool is an intuitive web dashboard that provides the public with a detailed analysis of the air quality in their area, as well as the potential resulting health effects. Users visit our dashboard and are presented with a map view of the United States and a set of filters. These filters allow the user to pick a particular address or region to learn more about the air quality in that location.
We use state-of-the-art machine learning techniques to make predictions about air quality for every location in the United States. Whenever a request for air quality information is made we use historical air quality data as well as up-to-the-hour live air quality readings from sensors around the country to make accurate predictions on the current air quality in any given location.
Our tool is used by a wide variety of people, including people trying to find a safe and healthy place to live, as well as policymakers trying to determine areas in need of assistance.
The front end is developed with HTML, CSS and JavaScript. Our software is served via a Python Flask back end which communicates data from our scikit-learn machine learning models. Our live data is retrieved from Purple Air Sensors.
Anthropocene Institute: Electricity Grid Planning Tool
The Anthropocene Institute is a non-governmental organization with the mission of utilizing science and technology to address the planet’s needs. It drives and facilitates innovation in clean energy to address the urgency of climate change. The Institute also supports start-ups and universities to develop emerging and disruptive energy technologies that are clean, safe and reliable.
The Anthropocene Institute is interested in the possibility of Small Modular Reactors (SMRs) as a viable option for new sources of power generation in comparison to coal and gas plants that have significantly worse emissions that affect the atmosphere.
SMRs, however, must be placed properly in power substations based on power generation, power outage occurrence rates, and power consumption demands. Traditionally this would be done entirely by hand, which is time-consuming, error prone, and makes it hard to quickly compare multiple placements.
Our Electricity Grid Planning Tool is a web dashboard that uses machine learning to automatically simulate SMR power generation and SMR placement. Our tool uses historical energy consumption data to accurately predict the costs and benefits of placing an SMR at a particular power substation.
Our web application provides an easy-to-use interface used by electricity grid planners seeking to better understand the cost and benefits for deploying SMRs. Users simply view specific substations, and our application provides them with statistics and recommendations relating to the cost-benefits analysis of deploying a SMR in that area. Users can also easily compare two locations to make an informed decision without wasting time with manual calculations.
The machine learning models were developed in Python with scikit- learn. The user interface is built on JavaScript, CSS and HTML with an Apache web server and Google Maps API.
Atomic Object: Stroodle: Learning Management System
Operating for over 20 years, Atomic Object is a software design and development consultancy based out of the Midwest cities of Ann Arbor, Grand Rapids and Chicago. Atomic Object has worked with over 175 clients and created over 250 applications across different industries, from tech startups to Fortune 500 companies.
Learning management systems are utilized by many educational institutions to administer, track and deliver course materials and student work. Popular offerings provide many features to manage and engage in course activities but fail to deliver them in a simple and intuitive application.
Our Stroodle: Learning Management System provides tools for students and instructors alike to participate in online courses while streamlining the user experience.
Students are provided a dashboard with a summary of important information for all their enrolled courses. They can access individual course pages to interact with material prepared by the instructor. Instructors manage their course by uploading documents, such as reading materials or assignments, for their students to view. They can also organize upcoming events and deadlines for their students on the course calendar.
Easy quiz creation is supported to assess students. The results of these quizzes are available in a student’s gradebook along with scores of other graded assignments. Students are sent push notifications when instructors send out announcements.
Our software supports all the key features of learning management systems without any of the confusing aspects, streamlining course management and decreasing wasted time.
The Stroodle front end is built using ReactJS and React Native. The back end is built using Node.js and Express.js, which communicates with a PostgreSQL database.
Auto-Owners Insurance: Yard Wars: Weathering the Storm
Auto-Owners Insurance is a Fortune 500 company headquartered in Lansing, Michigan with over 48,000 licensed insurance agents. Auto-Owners provides automotive, home, life, and business insurance to nearly 3 million placeholders in 26 states.
As an insurance company, it is important for Auto-Owners agents to be able to gather and analyze data regarding causes for claims. This helps them better anticipate which clients could be at risk of property damage and may need to submit a claim in the future.
Our Yard Wars: Weathering the Storm project is a virtual reality application where storms are simulated on virtual residences. Any damage caused by the storms is viewable in real time and in first person. This data is gathered, stored and displayed on an external website for analysis.
Users start by selecting the difficulty for the simulation, which changes the severity of the weather and number of trees that can be placed. Then, the user is tasked with placing trees around the property in order to replicate an existing or planned residence. Once they have finished placing trees, the storm simulation begins.
As the storm progresses, trees can fall and possibly cause damage to the home. Data about the simulation, including fallen trees and any damage caused, is sent to an external database for viewing and analysis on the website by Auto-Owners agents.
Our Yard Wars is an engaging game that provides Auto-Owners agents with useful information, assisting them with providing high quality service.
Our virtual reality software is developed in Unity and written in C#. We use a MySQL database to manage the data from the simulation, and it is communicated using PHP to the website, which is hosted on the same server.
Bosch: Hardware in the Loop (HIL) Vehicle Simulator
Bosch is a global engineering and technology company with roughly 395,000 employees worldwide. Founded in Germany in 1886, Bosch is the world’s leading supplier of automotive components.
Currently, Bosch uses a Hardware in the Loop Vehicle Simulator to correct errors with their software and perform tests. This system connects to specific hardware to simulate a vehicle on the road. This hardware, however, is quite costly and therefore only two systems are available to all Bosch engineers in North America. To resolve this, Bosch selected the PEAK PCAN USB Pro FD as a low-cost replacement for the previous hardware.
Our Hardware in the Loop system reimplements the core functionality of Bosch’s previous system on the PCAN hardware.
Our software allows the user to perform basic vehicle maneuvers and operations such as steering, braking, accelerating and more. The main functionality of our software system is adaptive cruise control (ACC), which users may engage and adjust at will.
The user interface is designed to allow anyone with driving experience to control the simulation with ease. A graph is displayable to show the signals being sent to and from the PEAK hardware with their corresponding values.
Above the controls for the vehicle is the dashboard. This displays the same basic information found in a real car including the current speed, rpm, fuel level and more.
Our software implements the same functionalities as the previous vehicle simulator but runs on much cheaper hardware, cutting costs and allowing more systems to be used concurrently.
The entire software system is written in Python 3. The front end is built using the open-source toolkit wxPython, while communication with the hardware is done using PCAN Basic API.
Delta Dental of Michigan, Ohio, and Indiana: Smart Benefit Plan Recommender Engine
Delta Dental is an insurance company that provides dental coverage to more than 80 million Americans, spanning across all 50 states, making them the largest dental care provider in the nation.
Delta Dental takes pride in tailoring benefit plans to their customers’ needs, whether they are a small business, a family or an individual. Before recommending the ideal benefit plan to a customer, Delta Dental underwriters must aggregate data from several sources and use their significant domain knowledge to properly recommend a benefit plan. Creating personalized dental plans for each customer takes a significant amount of time, as there are many factors that need to be considered.
Our Smart Benefit Plan Recommender Engine aids underwriters by automatically matching new customers with benefit plans that are used by similar customers, with no input from an underwriter. Potential customers use our website to answer a series of carefully crafted questions that are used to recommend a benefit plan.
Our Recommender Engine uses advanced machine learning algorithms to divide customers into groups that share many similarities. Each group is assigned an ideal benefit plan and when the system is given new data, it can easily provide a recommendation by mapping the new data to a group and its respective benefit plan.
Our system makes the insurance shopping experience less stressful for customers by allowing users to input their information through an easy-to-use interface and providing immediate benefit plan recommendations and links to help them enroll in the plan.
The front end of our system is written using Angular, while the back end is written in Python. The data is stored in a Snowflake database, and the clustering models were developed in Jupyter Notebook using the pandas and scikit-learn libraries.
Delta Dental of Michigan, Ohio, and Indiana: Microsoft Excel Data Extractor / Modeler
Serving more than 80 million Americans, Delta Dental is America’s leading provider of dental insurance. To provide quality service, the company must host and leverage complicated data.
Delta Dental maintains a significant number of sophisticated Excel spreadsheets for various purposes. Comprehending these spreadsheets requires significant industry expertise. Furthermore, extracting the industry knowledge from the spreadsheets to use in other applications often requires the development of computer programs designed specifically for a single workbook, which is both tedious and costly.
Our Microsoft Excel Data Extractor/Modeler is a web-based tool that removes the need for developing single-use computer programs for extracting Excel data. Our system imports existing Excel workbooks and, with minimal user input, can extract the relevant information as well as the data hierarchies present in an Excel spreadsheet.
Our web interface is designed to look and feel similar to Excel, but with specialized functionality that allows users to formally define data hierarchies and dependencies. These hierarchies can then be visualized to understand the structure of an Excel spreadsheet, as well as to reformat an existing Excel spreadsheet into an easier-to-understand form.
The web application and resulting data visualizations and formatting help enhance employee comprehension of complex spreadsheets and add extensibility to the existing data. Our tool removes the need for development of single-use programs, saving valuable time for Delta Dental employees.
Users can access the Data Extractor/Modeler through a web application that uses JavaScript and TypeScript within the MEAN stack framework. We use X-SpreadSheet for our user interface and ExcelJS to extract the data and formulas from Excel workbooks.
The Dow Chemical Company: Virtual Computer Service Enhancements
Headquartered in Midland, Michigan, Dow is a world leader in the innovation, creation, and distribution of specialty chemicals, advanced materials, and plastics.
As a large company with over 35,000 employees worldwide, Dow provides its employees with a service called Dow Virtual Computer, which is a virtual machine array. These virtual machines grant access to anyone with Dow credentials and a device with internet connection to a physical Dow work computer without the use of a VPN.
Currently, Dow has a personal structure for assigning virtual machines to their employees and clients on Microsoft Azure. This means that each user has access to their own virtual machine.
Dow is moving to a non-persistent state, meaning a single virtual machine’s resources can support multiple users, cutting down the required number of virtual machines running and therefore saving money. In order to do this, there are many manual processes that need to be managed.
Our Virtual Computer Service Enhancements software automates the previously manual processes associated with the transition to a non-persistent state by tagging user information, such as Dow identification number and login location, on each virtual machine. These tags help Dow with their billing process and the location assignment of the virtual machines.
Along with the automated tagging, our software generates weekly reports of users with multiple virtual machines. Dow’s support teams use this data in their transition to a non-persistent state.
Our automation process is built within the Microsoft AzureCloud Computing Platform and runs through Azure Automation Runbooks using PowerShell. The generated reports can be viewed through our mobile Microsoft Power App.
Evolutio: ERP Kids: Wildlife Conservation
Evolutio is a group of technology professionals that provides elegant solutions for complex business problems by leveraging technology in novel ways. Evolutio has partnered with the non-profit group Elephants, Rhinos and People (ERP) to give back to the global community through a variety of programs.
ERP is continually educating the world about wildlife preservation through its various charities and social outreach. Part of their goals for educating includes reaching children in the 6-13- year-old age demographic.
ERP Kids: Wildlife Conservation is a top-down role-playing game that allows the player to live as a wildlife ranger on the Dinokeng Reserve in South Africa. The game is offered on all mobile devices and designed to teach elementary and middle school children about the efforts of wildlife conservation rangers.
The game is divided into days and nights, in which the player has different tasks to complete based on the time of day. The player moves their character around an expansive two-dimensional map modeled after the Dinokeng Reserve in South Africa. The game provides the player with fun mini games to simulate tasks that would be completed by real rangers.
During the days, the player cares for animals, raises funds, and secures the reservation. During the nights, the player tracks animals and scares away poachers. The player progresses through the game’s narrative by completing tasks and improving the reservation.
The game informs the player of the importance of protecting wildlife and how this work is accomplished. The player also learns about the daily lives of rangers, the dangers faced by animals on reservations, and the ways in which people can help them.
Our game is developed using the Unity game engine and the C# programming language. Xcode and Unity are used to export the game to the iOS App Store and Google Play Store, respectively.
Ford Motor Company: Crowd Sourced EV Emergency Recharge
Ford Motor Company is a multinational automotive manufacturer headquartered in Dearborn, Michigan, with operations in over 125 countries and a worldwide workforce of 186,000 employees. Ford designs and manufactures a full line of cars, trucks, SUVs and electric vehicles under both the Ford and Lincoln brands.
Electric vehicles have recently experienced significant growth in popularity and technical advancement in the automobile industry.
A common concern of consumers who are hesitant to switch from a gas-powered vehicle to an electric vehicle is the possibility of running out of charge with no recharge station nearby.
Our Crowd-Sourced EV Emergency Recharge mobile application provides a platform for owners of electric vehicles to request assistance from other electric vehicle owners in the area if they run out of charge while driving.
After logging in to the application, a user is greeted by a Help Center screen, from which the user can submit an assistance request or view the open requests in their area.
Users submitting a request can input information pertaining to their vehicle, current location, and destination. The application notifies nearby users that a new request has been submitted.
Users wishing to aid stranded drivers can select a help request from the screen, view details, and select a price for which the user is willing to charge up the stranded driver’s car. Once the offer is accepted, the user is provided with the location, vehicle, and license plate information required to fulfill the request safely.
Our system addresses the concerns of potential customers, increasing sales and improving the reputation of electric vehicles.
Our mobile application is compatible with both iOS and Android. The front end is built with React Native and the back end is built using Swift and Java.
General Motors: Enhanced MISP User Interface
General Motors is an American multinational automotive company headquartered in Detroit, Michigan. GM is ranked 22nd on the Fortune 500 for total revenue and is the largest automobile manufacturer headquartered in the United States. For more than a quarter of a century, GM has integrated their OnStar in-vehicle safety and security service into millions of vehicles to become the most connected automaker in the world, with more than 22 million members.
Real-time, open-source threat intelligence is imperative in mitigating the risk of successful cyberattacks against an organization.The threat landscape is ever evolving. Consequently, GM’s security analysts need up-to-date information on all threats to ensure the organization’s safety.
The Malware Information Sharing Platform (MISP) is an open-source sharing platform dedicated to sharing threat intelligence. However, MISP’s user interface lacks several features, making it not very user-friendly or customizable.
Our Enhanced MISP User Interface provides for more customization within the MISP application and enriches existing functionalities.
The enhanced interface enables the removal and reordering of columns within the malware tables. Improved sorting functionality and search query filtering allow for more in-depth results from searches. The user interface boasts a more refined and intuitive design as well.
Using our software, GM cybersecurity analysts are able to create a version of MISP that is tailored to their specific needs and is easier to use, saving time and increasing productivity.
The front end of our software is developed using Bootstrap, a web application framework that utilizes CSS, JS, and HTML. Our back-end software functionality is implemented using PHP.
Herman Miller: Live Platform CAD Ingestion
With over 100 years of experience, Herman Miller is a globally recognized provider of furnishings and related technologies and services. Headquartered in Zeeland, Michigan, Herman Miller has been innovating new ways to design and analyze the places people work, learn and live.
Herman Miller has created Live Platform, a service which allows users to visualize and analyze their workspaces. Live Platform uses sensors placed on furniture to monitor occupancy and usage of different areas in the space. Live Platform previously used simple images to display workspace floorplans, which resulted in a loss of important data.
Our Live Platform CAD Ingestion software allows Herman Miller administrators to use computer-aided design (CAD) files to display these floorplans.
Our software allows for the supplying of sensor positions and the checking of hardware constraints. An algorithm is used to convert CAD files into a format readable by our web application.
The web application displays information on all floorplans. Each floorplan has a tenant owner, name, and is part of a campus and facility. Administrators can filter floorplans based on their attributes. The navigator tab allows for the finding and editing of floorplans displayed by the filter. Floorplans can also be added and removed. The viewer tab displays the layout of the selected floorplan. Users can toggle various floorplan information, such as sensor locations.
Our system automates the process of ingesting CAD files with sensor data into the Live Platform service, saving Herman Miller administrators time and increasing productivity.
The front end of the web application uses React, Redux and Bootstrap. The back end uses Node.js and Amazon Web Services, including Lambda, Simple Storage Service, and DynamoDB.
Lockheed Martin Space: SmartSat Satellite App Store
Lockheed Martin Space, headquartered in Denver, Colorado, is one of the largest space defense contractors in the world, employing over sixteen thousand people that develop an impressive range of products from satellites to space probes to missile defense systems.
Currently, Lockheed Martin Space is revolutionizing the way new software is written, tested and deployed to their diverse range of satellites through their SmartSat system. SmartSat provides a standardized format for software applications.
Lockheed Martin Space’s satellites can vary in a few key ways, specifically when it comes to computational hardware, software development kits, and operating systems. Because of this, certain satellites are incompatible with some specialized software.
Our SmartSat Satellite App Store is a web-based marketplace for browsing, uploading, and installing mission-ready applications to live satellites. The SmartSat App Store also includes rigorous software testing that automatically determines what software is compatible with which satellites.
Every new application uploaded to our app store is put through our automated compatibility testing to assess the Lockheed Martin satellites on which the software can be deployed. The test results are sent back to the App Store and displayed for application developers, saving them many hours of rigorous testing.
To ensure reliability, our App Store runs every new piece of software against every possible target hardware on real, physical devices. The ability to test on real hardware instead of simulation offers peace of mind to Lockheed Martin engineers.
The SmartSat Satellite App Store uses SmartSat Defined Services to perform the automated testing on the target hardware. SmartSat Satellites access these services through built-in file servers linked with the Flask back end. The front end of the SmartSat Satellite App Store is built with React.
Malleable Minds: Review Aggregator for Educational Programs
Malleable Minds is an early-stage startup company dedicated to allowing families to easily manage and share educational resources in order to connect them to global educational opportunities.
The company’s mission is to connect parents to the best possible educational programs in and around their local area and to receive continuous feedback from reputable sources.
With their many other duties, parents may not have the time or resources to research programs specific to their children’s needs. Our Review Aggregator for Educational Programs provides a solution to this by aggregating certified educator reviews in addition to reviews from other parents to make a one-stop shop for K-12 educational program browsing.
Parents can “like” programs, which will then be saved to their favorites page. They can also opt to receive notifications about a specific program. Educators and parents alike can rate programs, and then the ratings are displayed to all users browsing programs. Parents can also follow educators to see their ratings on other programs.
Understanding the experiences of previously enrolled children and the ability to compare educator ratings serves a critical function for parents in their search to provide enriching, educational experiences for their children. Our solution supports a growing community of educators and educational providers, allowing parents to make informed decisions based on relevant and current information.
Using a lightweight, contemporary technology stack, we built a responsive user interface so that users can access the application from any personal computer or mobile device. Our software is written in a combination of Python and React, and our system is deployed using Amazon Web Services.
Meijer: mHealthy: Healthy Eating Application
With over 250 supercenters, Meijer is one of the largest retailers in the United States based on 2020 revenue according to the National Retail Federation. Meijer offers a vast array of products ranging from home goods and furniture to pharmaceutical needs and groceries. They strive to assist shoppers to take care of the health and wellness of themselves and their families.
Many shoppers want to make healthy decisions when shopping but feel overwhelmed trying to do so. Finding healthy replacements can be a daunting task and may discourage shoppers from learning about healthier options or better diets. Educating shoppers and improving their experience is imperative to establishing lifelong customers.
Our mHealthy: Healthy Eating Application guides shoppers with nutritional advice according to their selected dietetic preferences. Several unique profiles can be created to allow shoppers to curate their shopping experience for family, friends and special events.
In a shopping session, a single profile is enabled to filter products that are accustomed to the shopper’s personalized preferences. As items are added into the cart, the application reviews these items and recommends additional products that better suit the shopper’s dietary preferences. Shoppers ultimately benefit from an easier and more informative shopping experience.
Our application encourages shoppers to eat healthy and provides an easy-to-use interface. Health-conscious shoppers are more likely to visit a Meijer supercenter thanks to our intuitive application.
Our mHealthy: Healthy Eating Application is both a web and a mobile application. The website is written using PHP, HTML, CSS and JavaScript, while the mobile application uses Java, Kotlin, XML and Android SDKs. SQL is used to manage and store the data.
Microsoft: Feedback Prompt for Ratings in Google Play Store
Microsoft is a multinational and industry-leading technology company best known for developing numerous operating systems, software, and online computing service platforms. Microsoft’s Intune Company Portal is a data and device management system that is used by tens of thousands of companies worldwide and requires individual employees to install its app on their devices. User feedback about Intune Company Portal provides insight that developers can use to create and maintain a premium user experience.
However, collecting and analyzing reviews is a tedious and time-consuming task, requiring manually searching dozens of review websites and plotting the data. Our Feedback Prompt for Ratings in the Google Play Store system improves the overall review collection and analysis through an easy-to-use in-app review prompt for users, and automated review collection and analysis for Microsoft engineers.
To facilitate easier and more representative review collection, we integrate an automated in-app review prompt that actively promotes user feedback for the Intune application on Android. Our system removes the need for users to navigate to the Play Store and makes leaving a review quick and easy. Our system employs a suite of web scrapers that search the internet for reviews of Intune and saves them for future analysis. We use natural language processing and machine learning to search for patterns in user reviews that might signify any bugs or issues and to determine sentiment towards the app. The results of the analysis are sent to Microsoft engineers using a Microsoft Teams Bot that delivers monthly analysis. The front end uses the Microsoft Bot Framework and Android Studio for the in-app prompt. The back end consists of a SQL Server database hosted on Azure. The analysis pipeline is hosted on Azure and utilizes Node.js web scrapers and Python scripts.
Mozilla Corporation: Improve High Contrast Mode for Firefox
Mozilla is a global, not-for-profit organization dedicated to improving the World Wide Web. They have an international community of developers who contribute to open-source software.
Mozilla’s most popular open-source project is Firefox, with over 210 million monthly active users. Mozilla’s goal as a company is to promote an open and inclusive internet. Part of this is making sure Firefox is usable by as many people as possible.
To achieve this goal, Firefox offers high contrast mode, which is a huge part of making the browser as accessible as possible. High contrast mode improves the visibility of Firefox’s user interface by changing the colors in order to maximize contrast. This is important as it allows people with visual impairments to be able to receive the entire Firefox experience.
While the high contrast mode currently in Firefox is an essential feature for making the browser more accessible, there is still more that can be done to make high contrast mode even better. Specifically, high contrast mode does not perfectly cover the entirety of Firefox’s user interface. This leads to confusion among users with visual impairments and an overall non-optimal user experience.
Our Improve High Contrast Mode project focuses on fixing many of these smaller issues to greatly improve the user experience of high contrast mode inside of Firefox. Our additions include things such as adding outlines to tab buttons, improving icon visibility and adjusting button colors to improve contrast.
In the screenshots to the right, one can see various pages of the Firefox browser using our Improve High Contrast Mode for Firefox project.
To implement our fixes, we edit and improve the already existing CSS, JavaScript, and HTML of Firefox. Additionally, some of our edits are written in C++.
MSU Federal Credit Union: Spaving: Giving based on Spending Habits
Founded in East Lansing in 1937, MSUFCU, also known as the MSU Federal Credit Union, provides various financial services to students, faculty, and staff at Michigan State University. With 21 branches, over 300,000 members, more than $6 billion in assets, and nearly 900 employees, it is the largest university-based credit union in the world.
MSUFCU provides a variety of financial education resources to its customers to ensure they are making the best monetary decisions possible. One of these resources, acquired by MSUFCU’s Credit Union Service Reseda Group, is a banking app known as Spave. Spave is a mobile application that allows users to donate a small amount to chosen charities every time they make a transaction.
Our Spaving: Giving based on Spending Habits project enhances the Spave mobile application with a recommender engine, providing users with recommendations for which charities the user should consider donating to, as well as alerting them to these recommendations.
Our most significant addition to this mobile app is a new recommendations page where the user can see the various suggested charities. These recommendations range from simple comparisons to other users, to recommending charities they may enjoy donating to based on how they spend their money.
Users can like or dislike recommendations to influence what is recommended to them. Recommended charities can be clicked on to present the user with further information regarding the selected charity.
Our front-end additions to Spave are built using Android Studio, Xcode, React Native and Typescript while the back end utilizes Express.js, Node.js, machine learning and Amazon Web Services.
PwC: Collaboration Bot for Microsoft Teams
PwC is a global leader in professional services, from tax and audit to technology and strategy consulting. PwC is a partnership of firms, consisting of over 250,000 employees operating under one brand, making it one of the largest professional services networks in the world.
Collaboration with clients is key to PwC’s business, and as such, their clients want to use newer and more ubiquitous technologies, such as Microsoft Teams and OneDrive.
Currently, collaboration requires external users to be invited into the PwC Microsoft environment and given access to Microsoft Teams and OneDrive securely. This requires significant work from an IT administrator who has to manually set this up, and in some cases seek out and get approval from the PwC clients. The entire process takes a significant amount of time and energy that could be better spent elsewhere.
Our Collaboration Bot for Microsoft Teams allows PwC employees to establish collaborative partnerships with external companies within the Microsoft environment with little to no work from an IT administrator.
Our software provides an easy-to-use interface within Microsoft Teams that allows PwC employees to easily send bulk invitations to external collaborators without having to manually approve each user and grant them access privileges.
Once invitations have been sent to external clients, our bot automates the entire approval process in the background, allowing PwC employees and IT administrators to spend their time on more significant issues.
The front end of our Collaboration Bot for Microsoft Teams is built using Node.js and ReactJS, which communicates with Microsoft Azure Active Directory and an Azure Database via Microsoft Graph API calls.
The Rocket Companies: ROCKY: Team Challenge Application
The Rocket Companies are made up of 15 publicly traded companies involved in many different industries, including mortgages, fintech, real estate, automotive and more. They employ over 26,000 team members and are committed to providing the best team member experience they can.
More and more companies are taking strides to improve the mental and physical wellbeing of their team members. Our ROCKY: Team Challenge Application improves the team member experience by promoting wellness with a fun and interactive web application.
ROCKY provides a platform for Rocket team members to compete in challenges against other team members to promote wellbeing, community and productivity. Challenges can measure any number of attributes such as steps walked, hours volunteered, and loans closed.
Challenges can be created as either team or individual challenges. To join a team challenge, team members can create a team, join an already existing team, or enter the challenge as free agents, who are assigned to teams at the start of the challenge. Team members who create a team can send invites out to other team members to join their team.
The challenges page has a list of all active challenges a team member can join and details about the challenges. This list can be filtered by the category of the challenge, the attribute being measured, or whether it is a team or individual challenge.
The user’s home page contains information about the challenges that user is currently participating in as well as any pending invitations to join a team challenge.
The front-end user interface of ROCKY is built in Angular. The back-end data is stored in an Amazon Aurora database with a .NET Core RESTful API to query the database.
Stellantis: Interactive Digital Assistant
Stellantis is a leading global automaker and a mobility provider headquartered in Amsterdam with operations in nearly 30 countries and a worldwide workforce of over 300,000 employees. Stellantis is guided by a clear mission: to provide freedom of movement through distinctive, appealing, affordable, and sustainable mobility solutions.
In today’s fast-paced world with responsibilities spanning multiple projects, staying up to date with the latest information is critical for Stellantis employees to make timely decisions and reassess priorities.
Our Interactive Digital Assistant is a web-based chatbot that allows Stellantis employees to ask general form questions in a life-like conversation about a wide variety of topics, including project statuses, issues or incidents, business applications, process contacts, etc. in a timely manner.
Our Interactive Digital Assistant first asks general questions of the employee to ensure they have the proper credentials to access the documents in which they are interested. Once the user’s identity has been confirmed, the user can ask any question they want.
Once a question has been received, our natural language processing algorithms parse the request and search multiple documents for the correct answer. Our Interactive Digital Assistant can provide answers to questions in less than a second, whereas manually searching all the relevant documents would take from minutes up to hours.
Our Interactive Digital Assistant helps employees find crucial information quickly, allowing them to spend their valuable time focusing on more important tasks.
Our front end is developed in CSS, HTML, JavaScript, and PHP. Our system utilizes the AWS Cloud Platform with Python. Our APIs are hosted by Google Cloud Platform and Drive IT.
TechSmith: Snagit Template Creator
Headquartered in Okemos, Michigan, TechSmith is a software company that develops screenshotting, screencast and video editing software. TechSmith’s products have over 73 million users worldwide. Snagit, one of the company’s flagship products, is a simple, but powerful screen capture and recording software that allows a user to quickly capture their screen, add additional content, and easily share with others.
Templates in Snagit are designs users can download to more easily create graphics instead of starting from scratch. Currently, Snagit templates can only be created using internal TechSmith tools, making it a challenge for partners or savvy customers to create their own templates.
Our TechSmith Snagit Template Creator platform provides a solution to this problem through an intuitive web application that allows users to easily create, download and share Snagit templates with others.
TechSmith Snagit Template Creator allows users to customize templates through a variety of graphics objects including drop zones, shapes and lines. Once the creation of a template is complete, users can download it, share it with colleagues or save it within the web application for later viewing and use.
When a user downloads a template, they can import the template into Snagit, where it can be utilized just like existing predefined Snagit templates.
The additional creative power and flexibility our system provides greatly enhances the Snagit experience and makes Snagit a more attractive option for those looking for a graphical editor.
TechSmith Snagit Template Creator uses Microsoft Azure for data storage and sign-in. The front end is built using ReactJS, and the back end is built using .NET Core. template into Snagit, where it can be utilized just like existing pre-defined Snagit templates.
United Airlines: Gate Hazard Geo-Mapping
United Airlines, Inc. is a leading American airline headquartered in Chicago, Illinois. In 2019, United and United Express operated more than 1.7 million flights carrying more than 162 million customers to their destinations safely.
Safety is United Airlines’ highest priority at every airport. A major component of that commitment to safety is being aware of the hazards faced by aircraft and operations staff at each gate.
United Airlines Safety & Compliance staff record data on the hazards faced at each airport gate. This information is then referenced by employees when they begin working at the gate to ensure both their own safety, and the safety of their aircraft during operation. United Airlines is exploring ways to further standardize this process and make the information more accessible.
Our Gate Hazard Geo-Mapping software allows staff to enter observed hazards into their mobile device, and automatically generates GPS placement data, as well as an interactive map visualization to allow seamless cataloging and sharing of gate-specific hazard information.
When United Airlines staff identify a hazard, they can open our application and select an airport and gate. They then mark the area where the hazard exists, record a title, description, the level of risk the hazard poses, and which departments should be aware of the hazard.
When other United Airlines employees visit a gate to work, they use our application to quickly familiarize themselves with the hazards at the gate to ensure they are adhering to the strict safety protocols.
The website front end is written in HTML, JavaScript, and CSS. The back end consists of a MySQL database and PHP code for both the API and data processing. The Android and iOS application is written in C# utilizing Xamarin for cross-platform development.
United Airlines: QA Audit Center
United Airlines is a major American airline company headquartered in Chicago, Illinois. Every day, United Airlines unites the world using the most comprehensive domestic and international network routes possible. This network connection requires United Airlines to hold themselves to the highest standards in safety and reliability, promote trust, and ensure flights are on schedule.
Within United Airlines, the Technical Operations Quality Assurance division plays a vital role in meeting United Airlines’ shared goals of efficiency, reliability, and safety.
To accomplish this, the United Airlines Quality Assurance team conducts live audits to ensure all equipment and services are maintained according to Federal Aviation Association (FAA) standards. Current audits require handwritten documents that are hard to keep track of and difficult to share.
Our Quality Assurance Audit Center Platform serves off-wing quality assurance auditors and provides them with a solution to have electronic access to audit documentation on their mobile devices. Our system converts a currently paper-and-pencil audit process to completely digital.
Our system recreates digital versions of forms used by auditors, which facilitates easier saving, editing, and submission of audit documentation. Additionally, our mobile applications allow real- time access to the camera to easily photograph and attach evidence to audit forms.
The time of auditors is valuable, and our system allows them to perform their duties in an efficient manner that reduces errors.
The front end of our application is written in Java for Android integration. The back end is hosted on Google Firebase. PHP performs government website scraping and stores updated data on the SQL Database hosted on the Michigan State University server.
Urban Science: Independent Repair Facility (IRF) Insights
Urban Science is a global data-driven company headquartered in Detroit that has provided tailored insights and solutions for the automotive industry since 1977. As a global company that has served every major automaker, Urban Science analyzes the market to pinpoint issues and propel success for their clients.
One of the most profitable components of automotive business is aftersales, the parts and services sold after a car is purchased. Independent Repair Facilities (IRFs) vastly outnumber dealers, dominating the aftersales sector. Without comprehensive information on IRFs, dealers miss an opportunity to increase customer retention and grow service revenue.
Our Independent Repair Facility Insights web application alleviates this gap of market knowledge through leveraging telematics data. Telematics systems in vehicles monitor a wide range of information including vehicle location, engine diagnostics and vehicle activity.
Using the telematics data, users are notified of what IRFs customers are selecting over the dealership for aftersales. This information is displayed on our web application and provides dealers with information such as IRF rankings and analysis on top- performing IRFs in their area of responsibility.
Users gain insights into the competitive landscape of aftersales through data visualizations and an interactive map. Leveraging clustering analysis and natural language processing, tailored solutions are generated for each dealer.
Our software enables the user to efficiently survey telematics data, explore critical components of IRFs, and utilize our data-driven solutions to better compete with IRFs and increase sales.
Our application is built using Angular, Typescript, HTML and CSS. The back end uses ASP.NET Core 5.0 APIs written in C# and is connected to an Azure SQL Database.
Vectorform: Smart Auto-Time Logging
Founded in 1999, Vectorform is headquartered in Detroit with seven offices across the globe. It is a company designed to help organizations move from an idea to an invention with digital products and hardware solutions. They combine a variety of technologies such as the Internet of Things, augmented or virtual reality, and other emergent systems to develop solutions for their clients.
Employees at Vectorform work on multiple projects at a given time using various tools and software development platforms. Keeping accurate totals of time spent on a given billable project is extremely important to both the company and their clients. Entering time statements manually is time-consuming and error-prone. Previous automatic time tracking systems fail to properly distinguish between different projects within one program.
Our Smart Auto-Time Logging system solves this issue by generating accurate and reliable time statements for software development and communications.
Users start the program before beginning their workday. The system generates accurate time statements throughout the workday by monitoring the programs in operation on the user’s computer and associating them to a project billing code through smart analysis.
Time statements are viewed through a web application to be confirmed before being sent to Vectorform’s billing department.
Our system automates the time-tracking process, eliminating the need for Vectorform employees to do it by hand, saving time and increasing the accuracy of the billing process.
The front end of our Smart Auto-Time Logging system is built using ReactJS, while the back end is implemented using C#. Both sides of the program interface with an SQLite database through the use of Node.js.
Volkswagen Group of America: Car-Net DriveView Social Competition App
Volkswagen Group of America is the North American operation headquarters and subsidiary of the Volkswagen Group, which facilitates the U.S. operations of many brands of vehicles. Car-Net was introduced by Volkswagen in 2013 with features that allow Volkswagen owners to easily access their vehicle with a mobile device, along with other features designed to improve the driving experience.
With nearly 5.25 million accidents per year in the United States, finding ways to encourage safe driving habits is an ongoing challenge. There is a need to encourage drivers to drive safely.
Our Car-Net DriveView Social Competition App is a social media mobile application that allows users to compete against each other by driving safely. Users who drive the safest are placed at the top of the leaderboards for other users to view.
Drivers achieve these safety scores by completing objectives relating to safe driving. With our application, users track their driving achievements by simply using their vehicle. A user’s safety score is calculated using these achievements as well as vehicle data.
To promote competition between users, leaderboards allow users to track how well they are doing compared to other drivers in their region.
Our application also allows users to filter who they are competing against. These filters include an overall distance driven by other drivers, other users’ overall safety scores and a radius of other drivers near them. In addition to those filters, users may also check how they stack up based on the current day, week, month or year to see how they have progressed against other users.
Our Android application is written in Java and uses Firebase for user authentication. The back end is written in Python with Flask libraries to communicate with the app through HTTP messages.
Whirlpool Corporation: AI Recipe Converter
Whirlpool Corporation, headquartered in Benton Harbor, Michigan, is the world’s leading major home appliance company with approximately $20 billion in annual sales and 75,000 employees. Whirlpool’s goal is to improve home life through the production of a variety of home appliances.
To this end, Whirlpool provides smart cooktops that automatically set the temperature and timers according to recipe instructions. Recipes must be formatted in a specific way to be used by these devices. Previously, it was the task of Whirlpool’s food scientists to convert recipes into the specialized format. However, manually converting these recipes is time-consuming.
Our AI Recipe Converter assists Whirlpool food scientists in this task by automatically converting recipes on cooking websites into the machine-understandable format. Using natural language processing, information such as ingredients, cooking temperatures, and cook times are extracted from recipes. The final automatically-generated recipe can be viewed on our web dashboard and exported to a Whirlpool smart cooktop.
Food scientists at Whirlpool upload a set of recipe URLs to our web application. These recipes are then converted by our software into the specialized format. Food scientists can then verify that the recipes have been converted correctly and make edits to the recipe. Finally, they can view all of their previously converted recipes.
The AI Recipe Converter saves Whirlpool time and accelerates the growth of their automated recipes library, ultimately alleviating the burden on food scientists and creating a better customer experience.
Our front-end web application is built with AngularJS while our back end utilizes MongoDB, Node.js, and Python. All components of the application are hosted on Amazon Web Services.
