Christopher Cao (USC), Josh Wolff (Stanford) - Devpost: https://devpost.com/software/cocao
In brainstorming ideas, we realized that one of the major pain points in healthcare is connecting underserved patients with providers. We were inspired by stories of undocumented immigrants who struggle to get the healthcare they need due to fears of deportation. Among many other, three of the main pain points for this population in healthcare are 1) they are not aware that many providers are willing to offer their services to undocumented immigrants and do not have to report any information to Immigration Security Enforcement, 2) language barriers in getting this information, and 3) an inability to trust others due to fears of deportation. The last point holds especially true in recent times with the Trump administration. Thus, we wanted to create a chatbot interface that enables undocumented immigrants and other users to get access to vital information regarding willing providers in the nearby area.
In order to achieve these three points, we created an iOS application to develop this chatbot interface. We hope to be able to develop an Android version in the future because Androids are much more common in the target population. We used Houndify’s API to develop custom commands to provide key information about healthcare providers in the nearby area. Commands such as “Where is the nearest hospital?” and “Are there doctors for undocumented immigrants?” return handpicked information about partners in the area. We hope to expand this functionality in the future by finding local partners to list on the application or finding doctors/nurses willing to provide services in their free time to patients who wish to stay anonymous.
Not only can users ask these questions in English, but by using a sliding menu, they can select one of any languages they might be more proficient in. This changes the language of the application completely, allowing them to speak to the chatbot in the language of their choice as well as receive responses in the same language. However, we were able to do this all while maintaining the full functionality of Houndify’s API and domains. We used the Google Cloud Translate API along with Apple’s built in speech to text features to make this possible. While with any kind of translation it’s difficult to capture certain nuance and intention, we would tackle this problem in the future by making more flexible, robust Houndify custom commands, so users can get the information they need even if the translation is imperfect.
While we mainly use this translated Houndify implementation for healthcare purposes, it’s important to note that this interface works with any of the domains offered by Houndify. Thus, this gives any developers hoping to develop with the Houndify API access to a huge population inaccessible before due to language barriers.