Fynd was inspired by the millions of citizens and residents of the United States that do not have a comprehensive understanding of the English language. Many immigrants and refugees face inconveniences every day due to the language barrier. In many of these cases, the result of that lack of communication can be as extreme as poverty, injury, or death. 44% of hospitals do not offer any linguistic or translation services. As a result, roughly 2.5% of all medical malpractice suits (including those involving death or serious injury) are due to insufficient language resources, putting the number of individuals and families that are directly and physically affected by this issue well into the thousands, and is likely higher due to the inability to follow the designated channels to file such a claim. Many immigrants are also restricted by their inability to interact properly with financial organizations and other institutions that, if used properly, would allow them to boost entrepreneurship, job creation, and abundant financial opportunities.
What it does
Fynd allows its users to easily access information on culturally-friendly resources in the area where they can receive care, services, and goods in their native language. With map and list functionality paired with feedback systems, Fynd looks to ensure that these services have high quality, cultural relevance, and convenient proximity to the user. Fynd generates its data with a combination of API plugins as well as the option for businesses to opt in and specify the languages that they have the resources to provide services within.
How we built it
For this hackathon, we decided to hone in on one particular use case--finding a doctor as an immigrant and/or refugee. First, we used a Python script to generate the database of the providers using the Better Doctor API by making requests with different parameters, aiming at mapping from an external id of uid to internal provider id and retrieving information from other tables to display on the front end. Backend is a flask server that gathers all the doctor information from a third party source, Better doctor. The data is hosted on a mysql server that stores all the reviews and reactions (thumbs up or thumbs down) that users give each doctor. All the backend runs on docker containers in a kubernetes cluster, making the app easy to deploy and scale.
Challenges we ran into
Since we had a team of mostly backend engineers, the front end development took longer than anticipated, and we learned quite a bit about react and general styling. Additionally, the translation task of allowing the user to search in their native language also turned out to be more complex than we’d initially scoped, so we had to leave that as a next step for Fynd. Lastly, we had to combine large pieces of code from many engineers in a very short amount of time.
Accomplishments that we're proud of
We are most proud of our ability to create a solution to an institutional problem that has been a persistent facet of the immigrant and refugee experience for years. In addition to developing a solution, we are also proud of our ability to demonstrate the data in a graphically simple and easily accessible fashion. We are very excited to deliver a full app. We have a fully staged, build-pipeline cloud hosted product (database, API, backend, and app).
What we learned
Language based communities of Immigrants can often be siloed in certain neighborhoods. Technology often exacerbates these feedback loops and information bubbles. However, we can easily use the business data collected from this app and collaborative filtering algorithms to guide outreach specialists with AI suggestion engines. This will mean targeting businesses nearby but outside these silos to expand their offerings of languages in the form of bilingual hires and signage/menus. We can break down the geographical barriers of these silos and expand services to these communities simultaneously
What's next for Fynd
Moving forward, we would like to explore other potential use cases beyond healthcare, add translation functionality so users can search in their native language, and apply a machine-learning based recommendation engine that would help suggest providers, institutions and businesses not only to help immigrants and refugees find culturally-friendly resources, but also help them explore their new country. Other use cases include, but are not limited to, financial institutions, childcare, legal aid, stores and restaurants. Generating a recommendation engine would involve building out and gathering the breadth and depth of data available through the platform.