MelaninRx

MelaninRx

Inspiration

Black people experience justified amounts of mistrust within the healthcare system due to past historical traumas like the Tuskegee trials, and the vicious gynecological experiments performed by Dr.Marion J. Sims. As Black women, we wanted to address this issue by providing black women with access to all of the resources they need to feel comfortable advocating for themselves in the healthcare system, be more informed about how symptoms show up specifically for black women, and better understand the biases in the medical system. We were inspired by our own experiences in the healthcare system and by several statistical facts that show how black people are underrepresented and disregarded in the medical system. For example, “research shows that African-American… children with autism are diagnosed at older ages than white children, giving them less of an opportunity for proper intervention treatment (npr.org).” Mistrust in the medical system amongst black people is a broad issue but we choose to focus on black women pregnancy’s as this has a troubled history and we might one day be affected by this ourselves.

What it does

We created a platform for Black women to empower their voices for better medical choices. We have three main aspects of our platform that address this. We have our AI chatbot, Our world cloud, and our profile analysis and comparison.

AI chatbot: This feature allows users to ask the chatbot any questions. It is currently connected to OpenAI but in the future, we hope to train the chatbot on data on Balck women's healthcare data in multiple fields like mental health, sexual reproduction, general, and pediatrics to name a few. That way the output from the chatbot would generate more targeted responses.

Our World Cloud: This word cloud examines articles about Black women's health regarding pregnancies, birth, and prenatal care. The word cloud displays the most significant words produced from analyzing the studies and weighs them, providing a view of the most prominent issues faced by Black women. The word cloud was then utilized by the AI Chatbot to help develop its scripts.

Health Analysis: Health Analysis and Profile Comparison: This feature enables patients to input their medical history and compare it with data from other Black women, as recorded by the CDC. Its objective is to empower Black women to identify others within the dataset sharing similar medical backgrounds, particularly in relation to pregnancy. Such comparisons aim to provide insights into how specific medical histories may influence pregnancy outcomes. Ultimately, this tool is designed to facilitate more focused research on the symptoms experienced by Black women, improving diagnostic accuracy and treatment efficacy for this demographic.

How we built it

We built it using these tech stacks: AI Chatbot: Open AI API, Express, Cors, React, Next, Node.js

Word Cloud: Python, React, Apache Echarts, MatPlotLib, Natural Language Toolkit, Pandas, scikit-learn

Health Analysis: Python, Flask, React, HTML/CSS, JavaScript, Pandas

Challenges we ran into

AI Chatbot: I was utilizing an old, discontinued version of Open AI at first so faced numerous errors before reading the documentation more in-depth. It was my first time using APIs, endpoints, and frontend/backend communication so experienced a significant learning curve, however, ended in success.

Word Cloud: The first library and method I used to create my word cloud was inconsistent with the developmental server and project React version. Due to this, I had to pivot and recreate my code using the Apache escharts library. This was very confusing and had its own set of complications but after much trial and error, I was able to get my component running to some extent.

Health Analysis and Comparison Analysis: Utilized CDC Wonder to create a sample data set. It was difficult to learn how to work within this system because I couldn’t figure out how to group by more than five attributes. I also have never created a full stack so it was challenging to navigate how to connect the HTML (front-end) and the python (back-end) to each other. That took a while to get running smoothly so it limited the amount of time I had to refine my logic on how I measure similarity between the user input and the records data.

Accomplishments that we're proud of

We’re proud of our in-depth research and analysis of Black studies and datasets to ensure we have a strong understanding of the problem we’re studying. We’re proud of our ability to develop a project with a front-end and back-end, developing our skills in React, next, apis, flask, and other skills. We’re proud of our ability to implement APIs for the first time through the OpenAI API. We’re proud of our ability to engage with the Palantir representatives and learn from their skills and experiences. We’re proud of our development as a team technically and socially, developing communication and collaboration skills.

What we learned

Our team learned a lot during the creation of this project. Every member gained a skill they had not known before, from implementing an API to utilizing Github as a team. An important skill we learned was how to work as a technical team. It was very difficult to create separate components that would eventually come together as one and through active communication and help, we were able to make our project compatible across each of our parts.

What's next for MelaninRx

Given more time we would like to develop MelaninRx further by: AI Chatbot can be trained with Black datasets to have a stronger understanding of Black experiences The dataset for scripts can be expanded using more keywords to have a wider variety of potential scripts for patients Word Cloud can use more research papers and data sets to gain a better understanding of key issues within Black medical experiences Health analysis profile comparison can incorporate additional attributes like medical history and stress levels