American-born Chinese (i.e. ABC’s) in their early 20’s aim to practice Mandarin speaking but feel anxious due to shyness or a feeling of cultural disconnect.

We aim to assist them in overcoming these barriers. Our approach involves using generative AI to foster conversations and create bonding with their speaking partners, ranging from family elders, new friends, or even romantic partners.

Ultimately, our goal is to empower these young adults to speak Mandarin comfortably and feel connected with their Chinese heritage, thereby strengthening community bonds among fellow Chinese speakers.

Sparq is an app that aims to reshape the way you connect and build lasting bonds. For people whose friends are their chosen family, Sparq is your personalized guide to discovering and nurturing friendships that align with your values and interests. 

Sparq guides you through a journey of self-discovery, helping you understand your expectations in a friendship and how to foresee and navigate change. By answering thought-provoking questions and engaging in activities, you'll set the stage for connections that truly resonate.

Our app provides an arena for friends to converse honestly and become closer, helping build deeper understanding of each other.  We believe in making interactions purposeful, enjoyable, and eye opening. 

Join us at Sparq, where the journey to genuine connections begins. Let's redefine friendship together! 

What’s Sparq?

Basically, our core assumptions going into this project are that people tend to avoid growth conversations unless they are forced to - by a conflict, or an external experience that raises unavoidable friction. Our underlying hypothesis is that people can learn to take preventative care of their friendships instead of waiting for a situation that needs a cure. 

We envision a tool that helps people foresee what issues might arise within a friendship and proactively address them. This could look like: thinking about the successes and failures of past friendships or relationships, and practicing projecting them into future hypotheticals. 
Part of this process, of course, is understanding and proactively communicating your own needs within a friendship. That will also be an integral part of our platform. 

The details will be refined through a rigorous process of UX research and iterative design.

In today's dynamic and rapidly evolving job market, navigating career transitions can be a daunting task for individuals across all cultural, economic, and social backgrounds, even more so for folks with limited access to learning resources and mentorship. CareerNav. was developed with a vision of enabling equitable access to career opportunities for everyone, beyond their measures. We aim for our platform to serve as a beacon of hope for economic mobility, providing every individual with equitable access to boundless career opportunities, transcending limitations and leveling the playing field for all.

CareerNav. is a tool that lets individuals, intending to transition from one career domain to another, identify transferable skills and skill-gaps in their profile. The goal is to help the user with a personalized learning map for the skills that matter in the industry. Users first update their background and career goals. The tool takes these into consideration and evaluates the transferable skills they already have, benchmarking them against industry trends. The tool then does a skill gap analysis and recommends a study plan for each skill gap. The user is also provided with an explanation for the recommended learning path.

Motivation

It is no surprise that understanding oneself is a necessity for leading a fulfilling and purposeful life. Research has consistently shown that self-awareness is linked to increased resilience, improved relationships, and greater overall well-being. By fostering a deeper connection with oneself, individuals are better equipped to navigate life's challenges, pursue their passions, and cultivate meaningful connections with others.

Problem

Many traditional self-assessment tools such as online tests offer simplistic assessments that fail to capture the complexity and nuances of human experiences. Additionally, they lack the interactive and engaging features that motivate users to actively participate in the assessment process. Without meaningful engagement, users may quickly lose interest, leading to inaccurate or incomplete results. Our project aims to help users understand themselves better in an engaging manner. We want to bridge the gap between traditional self-assessment methods and engaging self-understanding experiences.

Solution

Relatable gives users a space to explore themselves by engaging with their favorite movies in an introspective way. We have uniquely integrated cinema and psychology, allowing users to gain valuable insights into their own beliefs, strengths, desires and flaws. By analyzing character traits, thematic elements, and narrative arcs, users can uncover parallels between the stories on screen and their own lived experiences. 

Harnessing the exponential capabilities of Large Language Models (LLMs), we've curated an extensive database of movies enriched with their psychological characteristics. As users select their preferred movies, our unique algorithm finds patterns within these characteristics, providing insightful reflections on their psychological traits. Additionally, users can receive tailored movie recommendations based on the specific characteristics they resonate with most deeply.

After using Reelatable, users report feeling enlightened and empowered, gaining new perspectives on their thoughts, feelings, and behaviors.

Quantaide is an AI-powered learning platform designed to assist users in research methodology and survey design. By harnessing AI guidance, Quantaide enables users to craft precise and impactful surveys tailored to their specific needs. Through accessible learning modules and insightful evaluations, users are empowered to navigate the complexities of research methodology with confidence. Furthermore, Quantaide tailors the learning experience to the user's domain of interest, providing guided modules that ensure active participation while maintaining the "human-in-the-loop".

Key Features

• Learn Research Methods & Survey Design: Learning modules help users understand what goes into building precise domain specific surveys.
• Design Domain-specific Surveys: Upload any document that helps Quantaide learn about your domain.
• AI-Empowered Evaluation: Quantaide-AI tells users where they can improve in each module with a score and suggestions for improvement.
• Human-in-the-Loop Approach: Quantaide ensures you're in control of your research journey, with AI serving as a supportive guide rather than taking over your work.

It is hard to stay on top of all things that one is interested in. News podcasts seem like a good solution to catch-up on things while one is going about their day, however, no single podcast covers all topics of interest, tuned to individual preferences. Enter Echotune. We take user preferences such as topics they are interested, questions they would like answered, the duration of their commute or free time etc. Based on these we scrape the internet including social media channels to create an audio output summary of the desired length that they can use to catch up on things that the users are interested in.

Challenge

Recycling materials like plastic and fabric is essential for mitigating climate change and minimizing our carbon footprint. However, non-recyclable materials that end up in landfills contribute to severe environmental issues, such as water pollution and wildlife harm. Recycling centers play a pivotal role but struggle with the effective identification and sorting of various materials.

Textile production significantly impacts the environment, responsible for 10% of global CO2 emissions and 16-35% of the microplastics in oceans. In the U.S., less than 15% of textiles are reused or recycled, with less than 1% actually reprocessed due to their complex compositions. For instance, a garment labeled "100% cotton" may contain other fibers like polyester, complicating recycling processes. This issue underscores the challenge manufacturers face in recycling textiles, which involves sorting, purifying, and reprocessing into new products.

Approach

This significant industry gap motivated us to address this challenge by partnering with a startup that focuses on accurately sorting garments. We evaluated three companies: Sortile based in New York City; Recycleye in London, UK; and Refiberd in Oakland, CA, all of which employ advanced technology to identify fabric compositions. We chose to continue with Refiberd due to its proximity and innovative technology.

The Refiberd Tag Reader was developed as a solution to streamline the process of labeling new data, enabling Refiberd to train their machine learning model more effectively. This application captures an image of a garment tag and displays the fabric composition in the format required by the company's processes. Once the user verifies that the tag has been read correctly, the information is saved in a database. While the application interface appears simple, most of its processing occurs in the backend where a model processes the image and interprets the results, and the backend saves this information in a database. Additionally, as per client requirements, this app integrates with Amazon Web Services (AWS), ensuring a structured and robust architecture for the solution.

A student toolkit for understanding and evaluating the emergency mass notification alert system at UC Berkeley and/or at your university. This toolkit was developed by centering those of us in the most need, as described by Angela Glover Blackwell’s “The Curb-Cut Effect.” It is the hope of the research team that university students, regardless of their experience and background, impacted by their school’s application of the Clery Act can better understand the system and what may need to be brought into student community discourse or, with deepened understanding, brought into student community discourse. The intent of this project is to disrupt data access power structures and invite all people to have the resources and information to improve their lives and the systems that impact the quality of their lives. A range of resources were applied or developed and then tested to serve as a type of toolkit or “roadmap” to evaluate university emergency notification systems for students and by students.

Motivation 

Over the last 30+ years, little empirical data has been gathered regarding the effectiveness of the Clery Act, enacted in 1990. One particular area of interest to our project is the student perceptions of the emergency notifications that UC Berkeley sends out, in part, to comply with a 2007 update to the Clery Act which was motivated by the Virginia Tech shooting.

In September 2020, UC Berkeley was fined $1.35 million by the Department of Education for violating the Clery Act, including "failure to issue emergency notifications" and "failure to issue timely warnings in accordance with federal regulations". On February 9, 2024, it became clear that UC Berkeley's WarnMe alert system has weaknesses that represent a serious problem and require robust changes and new policies.

Our project/toolkit intends to inform people about campus safety, the Clery Act, and possible ways to stay safe. It explains detailed information and anyone can use the roadmap as a guide with reduced risk for targeted harassment and violence.

Solutions

Our capstone study on UC Berkeley's emergency notification system, WarnMe, uncovered valuable insights for improving campus safety communications. Through analysis, we found that location details and scene descriptions in WarnMe messages often lack the clarity needed to be fully effective. Interviews and a diary study with students highlighted a gap between the perceived and actual security provided by these alerts, leading to unnecessary anxiety and confusion. Also, how users conceptualize the administrators and authors of WarnMe messages influences their engagement with the system and their sentiments towards it. Sentiment analysis further revealed varied emotional responses, and disclosed the need for sensitive communication. These findings demonstrate the necessity of refining the emergency communication strategies to enhance safety and trust within the campus community.

Deliverables

• A student toolkit to understand and evaluate the emergency mass notification alert system at UC Berkeley and/or at your university including a comprehensive guide on conducting UXR interviews and diary studies.
• A report detailing our foundational research findings, such as coding WarnMe, sentiment analysis, and detailed user-experience research, along with suggestions for future improvement and development of the toolkit.
• GitHub repositories of Sentiment Analysis and Inter-Annotator Agreement Calculation provide a detailed explanation of how the analyses were conducted. 

Thank You’s and Appreciations

Our research team would like to thank the community of people that supported our research. We are grateful for the opportunity to learn and grow together as a mini-capstone community. Our appreciation is extended to our project advisors, Dr. Niloufar Salehi + Dr. Alex Hanna, the entire School of Information staff and faculty, The Daily Californian staff, MIMS Class of 2024, as well as our personal support systems, including friends, family and loved ones.

There is an increasing recognition that users should be able to exert greater control over their data – in law as well as broader civil society. However, current technical architectures underpinning platform ecosystems may not be designed to address the requirements of decentralized data ownership. 

Solid, also known as "Social linked data," is a set of protocols and tools aimed at constructing decentralized social platforms adhering to Linked Data principles with a vision to build a decentralized social network. At its core, Solid enables individuals and communities to securely store their data in decentralized containers known as Personal Online Datastores (PODS). With PODS, users have full control over the access permissions granted to both individuals and applications seeking to interact with their data.  

Solid applications currently lack readiness for widespread adoption due to significant usability and experience challenges, particularly for non-technical users. In response, PodBox, built on top of the Solid framework, offers a one-stop dashboard with a primary goal to simplify and streamline decentralized data management, making it accessible to all. By abstracting the complexities of the underlying protocol, PodBox aims to create a user-friendly interface for managing resources and permissions within one's POD, providing a unified and straightforward experience.

PodBox Features

• Data Ownership - Store all your data on the internet in one place and you control what to do with it!
• Interoperability - Seamless data sharing across all apps and services. Easily carry your data as you switch from one app to another
• Secure Sharing - You decide who—be it a person, group, or app—has access to resources in your Pod 

Problem Statement

This research project is guided by the question of how photo sharing on social media platforms after an art museum visit by visitors can influence their experiences during and after the visit. The diagram below describes the transition of our inquiries starting with this question and involves the examination of existing literature, semi-structured interviews, and data collection at the intersection of museum visitor experience, social psychology, and social media behaviors.

During the brainstorming phase of this research, we conducted a pilot observational study at the  Berkeley Art Museum and Pacific Film Archive (BAMPFA). The results of this study enriched our understanding of the social aspect of art museum experiences. They also guided our further literature review in the direction of art museum visitors’ photo-taking activities, their photo-sharing behavior on social media platforms, and the social psychology theories behind their online expressions and interactions regarding art museum experiences.

In particular, we are interested in addressing the following questions:

• How are visitors taking pictures and sharing photos on social media after a museum visit?
• How are visitors experiencing the in-person visit mediated by their photo-taking? How does social media sharing affect the museum experience?
• How do people experience social relations while experiencing museums? How are social interactions mediated by photo-taking and social media sharing?

Based on the findings across all methods, we hope to provide design recommendations for initiatives, engagements, and integration that small-scale art museums can adopt to attract younger audiences, increase engagement during and post-visit, and enhance the experiences of visitors given the values that these institutions uphold. 

We would like to discuss how the meanings of an art museum visit have evolved in the era of social media. For instance, they may no longer be simply about learning and viewing things. Is it necessary for art museums to adapt to these changes? If so, how might we make design recommendations to help art museums achieve this without hurting their fundamental values and purposes?

Key Activities

1. Observational Studies: We have and will continue to conduct observational studies primarily at the BAMPFA as well as other art museums to identify the visitors’ behaviors and engagements with their digital devices and their photographing patterns. During our pilot observational study, we have also considered the differences in navigation and engagement patterns between solo visitors and those visiting in pairs and groups.
2. Social Media Qualitative Coding: We will passively collect publicly available data (primarily photos, captions, and comments) from Instagram posts that have been tagged or associated with the museums under investigation. Then, we will carry out a detailed coding exercise to identify themes and patterns in the subject matter being photographed, the mood and tone of the photos and the captions, the level of interactions in the comments, the narratives associated with the overall post, and other aspects of the post.
3. Semi-Structured Interviews: To further investigate the motivations and impact of photographing and photo sharing post-visit, we will recruit and conduct semi-structured interviews with art museum-goers and Instagram users to uncover their perspectives and tendencies.
    

About Kipitup

Kipitup is an Apple Watch app that provides fitness advice for any athlete, no matter where they are on their journey. It delivers fitness advice that incorporates both the wearer's biosensory feedback and industry leading best practices, ensuring that advice is always specific to what would be best for the wearer. Kipitup is able to provide specific steps for improvement and feedback for 'cardio training' and ‘marathon preparation’.

User interactivity

Users will interact with Kipitup using by asking natural-language questions to their Apple Watch. After Kipitup takes into account the user's biofeedback and trends, it will reply with customized user-specific results. Those results are catalouged in the app so users can see historical advice and instructional plans.

How Kipitup works

Users vocally ask a fitness-related question using the Kipitup app in their Apple Watch. Kipitup converts their question using speech-to-text. The user’s question is directed to a large language model (LLM) we have trained to provide industry-leading fitness advice. In parallel, a biosensory ‘snapshot’ of the user is taken using the Apple Watch. Kipitup sends the biodata to the LLM (stored in a AWS Lambda instance) which processes it and converts it into a machine-readable string. The processed biosensory data is also directed to the LLM to provide additional context about the user. Finally, Kipitup generates fitness advice based on existing best practices which are tailored to the user’s biodata, ensuring that recommendations are always unique and customized to the user. That generated fitness advice is then sent back to the user via email.

Download Kipitup

Kipitup is currently available as a locally downloaded file. Instructions for how to load Kipitup to your Apple Watch are in the 'Final Paper' document. The download links below access a Google Drive in which these documents can be downloaded directly.

Kipitup App (.zip):          [Download]

Final Paper (.docx):        [Download]

Introduction

Sparking connection and conversation within your local community is difficult. With the innovation and integration of complex socio-technical systems in our immediate environments, people are no longer as motivated to reach out to their surroundings to express gratitude and excitement for the community around them. Creating neighborhood relationships is also no longer as simple as it once was - fear of digital deception and inauthenticity, risks associated with identity theft and online doxxing, and the competing interests of other social networks prevent the growth of unassuming, modest connections within neighborhoods. 

Through detailed ideation, concept testing, and design prototyping– our capstone project presents Collective. This product aspires to bring back dialogue amongst your community by emphasizing positivity and growth in local neighborhoods. With the foundation built by social psychology, computer-mediated communication, and digital trust-building principles, we have created a product designed to engage and excite people about their local neighborhoods. 

Problem Statement

How can we build on reasoning from literature and theories in social psychology and human-computer interaction to put forth a user-centered product that weaves together positive digital expression, establishes privacy and safety, and builds trust online for the creation of an experience of collective good?

The Problem

Research indicates that children globally are increasingly disconnected from nature, spending more time indoors. This lack of connection with nature is evident in reduced species literacy and a cultural shift away from nature-focused content in media and popular culture. Factors contributing to this disconnect include technological distractions, reduced emphasis on outdoor play in schools, and urbanization limiting access to green spaces. This lack of nature connection is concerning because it's linked to reduced well-being and less engagement in sustainable behaviors.

Nature-connected individuals tend to be happier and more likely to support environmental conservation. Given the rapid decline in biodiversity due to human activities, fostering nature connection in children is crucial for promoting sustainability. What’s good is that nature connectedness can be increased, and outdoor play can help get children engaged with nature and connected. 

Our Goal

Our capstone team, Playscape, aims to cultivate children’s connection with nature through an outdoor play experience.

Approach

Our approach features an iterative design process and includes the following activities:

• Literature review to identify research on outdoor play, its benefits, and its challenges
• Interviews with parents on children’s relationship with nature and play
• Interviews with experts on biodiversity, nature conservation, and science education
• Competitive analysis on current digital approaches to outdoor play and nature education
• Observational study
• Usability testing

Our Solution

Playscape is an outdoor play system that invites children to learn more about local biodiversity while engaging in playful learning. Playscape consists of an outdoor interactive installation and an accompanying digital application, focused on educational and sensory experiences.

Young adults in the United States below the age of thirty-four remain underserved by available political information sources, report more negative attitudes toward politics, and engage less in politics relative to other age groups.

We studied University of California, Berkeley students’ experiences with local political information and engagement using a combination of qualitative and quantitative research methods. Informed by opportunities grounded in the study of local student perspectives, we designed a mobile application-"Magenta" for young adults that provides them with highly accessible local political information.

Magenta directly addresses the challenges of community and political engagement found in our analyses starting with understanding the diverse needs of the users. Our research-driven personas, and personality quiz provide unique insights and allow us to tailor the user experience and simplify complex issues. 

Magenta builds on previous studies by reducing the negativity surrounding politics, removing barriers to information, which have been associated with discouraging participation.  Additionally, Magenta provides a novel approach by providing a platform that connects individuals with the local community which has been linked to increased sense of community and political engagement. Magenta’s innovative approach has the potential to bridge the knowledge gap, empowering citizens and making democracy fun.