One of the popular tools on the Pitchbook platform is their search engine, that allows investors to find what they are looking for from nearly a million records of possible investments. Currently their search engine offers an overwhelming number of search results without any meaningful order, and leaves little room for discovery. Our capstone project attempts to fix that by introducing a relevance factor to rank the search results, and by allowing the user to discover possible investments similar to previous successful investments, made either by the investor himself or by others. This makes search more meaningful and actionable for Pitchbook’s clients.

Thousands of travelers and homeowners use peer-to-peer homestay services like Airbnb every day. However, there is a major problem existing. The home management systems are not efficient, secure, and user-friendly. Most houses manage traveler’s entry-exit through conventional locks. When people book accommodations, they have to go through the hassle of collecting, using, and returning the key. Often, travelers reach the house during odd hours which becomes inconvenient for both parties. Also, traditional lock and key systems can lead to fraud. Our project solves this real-world problem by implementing a secure and user-friendly cloud-based IoT lock and key system.

Ensuring the health of shipments is of paramount importance to a Fortune 500 global logistics services company. Project Superbug assists Expeditors in doing exactly that by creating an automated reporting system that provides a data architecture supporting business intelligence and analytics. Precious man-hours are saved through this system, thereby providing analysts with more time to examine shipments. This brings great value to the health care client sector by enabling diagnosis with rapid turnaround times, which in turn creates faster treatment decisions, leading to better patient experience

In the telecom industry, the process of resolving service tickets is cumbersome and time-consuming. The support staff has to manually categorize these tickets and quickly find resolutions. We present “Ticket Genie,” a predictive analytics framework that expedites this process of finding resolutions. The framework automatically predicts resolution categories of new tickets and clusters similar tickets together using predictive machine learning techniques with up to 90% accuracies. Being deployed as a web service, Ticket Genie can be integrated into various ticket management systems with minimal efforts. With this expedited processing, we hope to achieve faster resolution leading to greater customer satisfaction.

The City of Seattle's Demographic Disaggregation Task Force (DDTF) is a cross-departmental group charged by City Council to change the way the City collects race/ethnicity data. With this more accurate information, City services can be targeted more equitably. Our project is a pilot of demographic disaggregation at the Utilities Discount Program (UDP), which provides discounted electric utility rates for low-income households. Through organizational research, observations of staff, and interviews with key UDP players, we came to understand the current practices around race/ethnicity data. From here, we made recommendations on how to integrate the disaggregated race/ethnicity data practices into UDP's framework.

Our team is working on improving the earphone paradigm for better usability and customizability by creating a modular device that will separate the ear attachment from the audio driver along with the rest of the assembly in addition to modifying the software behavior. Allowing users to customize what has historically been a one-piece device will enable them to selectively improve one aspect of their earphones without sacrificing any others. The modular design will also enable the customization of style and color, allowing the headset to double as functional jewelry and providing flexibility to the ever changing modern man.

The Financial Inclusion Insights (FII) team at the Gates Foundation is responsible for performing data quality checks and analysis on survey data collected from eight countries in Africa and Asia. The survey data is related to usage of mobile money and digital financial services (DFS) in these countries. The current process is manual and tedious and involves multiple people working simultaneously to perform data analysis. Our solution solves this issue by automating the manual process and generates customizable data analysis reports. The result is a robust, modularized and highly customizable framework which will make it easy for FII to explore the “what”, “how” and “why” of demand­ side trends in mobile money and other digital financial services (DFS). This helps FII model, construct, and perform better research to understand the impact of DFS and promotes adoption of these services as they are revised using the research insights.

In the realm of online content retrieval, understanding user intentions based on search behaviors is essential to connecting customers to the content they seek. Adobe Systems recently acquired Fotolia, a digital asset supplier, to more efficiently provide their customers with quality images and videos from inside their ecosystem. Working with Adobe Stock, we accessed all of their query and content engagement data collected to date. In this formative analysis, we applied natural language processing to search terms. Results were paired with multiple metrics of user interaction associated with Adobe’s content. These data were grouped, or clustered, to reveal hidden layers of similarity across queries. These clusters, representing similar user intentions, help identify Adobe’s underperforming segments of customer interests and behaviors. Once identified, treatments such as user interface modifications, search algorithm changes, or query refinement suggestions can be targeted to queries in the same cluster. This will enhance the Adobe Stock user experience¬ increasing customer retention, satisfaction, and spending.

The future of the country lies in the hands of the children. If the child doesn’t get the right start, he is 60% more likely to never go to college and 70% more likely to be arrested for a violent crime. However, good quality of early learning education can change these statistics. The Childcare Quality & Early Learning Center for Research and Professional Development (CQEL) is the University of Washington’s effort to measure and improve the quality of early learning environments for children. CQEL is responsible for conducting, onsite evaluations of child care providers located throughout the state. The process of rating early education sites was being executed using a complex and cumbersome Excel spreadsheet until Data Wanderers pitched in. The team transformed the IRC Database solution from Excel to a robust SQL relational database thereby catering the problem of data security, data losses, multi-user access and facilitates standard and adhoc reports to analyze the performance of various sites.