Alexa MD

Inspiration

The current healthcare system in Canada is one in which many people try to utilize services but because of the large masses, the system can become clogged. As a result, many people wait for for weeks or even months for a doctor's appointment. People wait for hours at the office just for a simple diagnosis. Other online solutions are not as convenient or trustworthy (for example, websites like WebMD always give you an unrealistic disease). We decided that we would fix this. We Alexa MD to be a convenient, accessible, and modern way of receiving a diagnosis. More than 3 million Canadians currently have access to a smart device such as Alexa, so people will be able to receive a diagnosis instantly. Also, using Alexa MD, people can find out about any potentially dangerous diseases before symptoms become too severe. We believe this technology has the potential to help millions receive a quick diagnosis at home and become educated about any harmful diseases that they may have before it gets more severe.

What it does

Alexa MD first takes user input. It asks the user various questions, such as what symptoms they are experiencing / how severe they are. Alexa MD is based of a data mining from various professional medical organizations. Using complex mathematical algorithms, it can accurately discern the condition that a user has based off what symptoms. This means that even if a user doesn't provide all symptoms of a disease, the algorithm can still correctly predict the condition. The keywords used by Alexa MD are programmed to be everyday words that would be in a typical person's vocabulary as well.

How I built it

We built Alexa MD as an Alexa Skill, an add-on to Amazon smart home products such as the Echo and Echo dot. Using Voiceflow, we linked Alexa to our database of medical conditions and symptoms, and using an algorithm programmed in javascript, we predicted what disease the user could have. We also used a machine learning algorithm to find a cure given a disease, which is much more efficient than researching each individual cure.

Challenges I ran into

The biggest challenge we ran into was connecting to the Echo Dot. Because the internet was not very fast, we couldn't test our product on the actual Echo until the second day. We had to work around this problem and find a way to use an Alexa Simulator. Also, because we were all new to the Alexa API and Voiceflow, it took some learning to get used to it. In the end, hours of hard work paid off and we completed our product!

Accomplishments that I'm proud of

We were super proud of building our own data set and using AI to help construct the data. Gathering the information and weighting it so the algorithm could use it was quite difficult, but we ended up finishing it. We are also proud of learning how to program Alexa skills. Learning a new API can be very difficult, but we pulled through and conquered the challenge. Overall, we learned many things over this hackathon and we are extremely glad we participated.

What I learned

We learned how to program Alexa skills using the Alexa API. We think this will be useful in the future, as smart home products are becoming more and more popular. Devices such as Echo will be used more and more for essential functions. We also learned more about machine learning and gathering medical information the internet.

What's next for Alexa MD

In the future, Alexa MD can be expanded to include more diseases and symptoms. It could also provide local information on how to get to the nearest doctor and what prescription drugs are available in the area. We could also incorporate doctor input on the data to help expand the database, making Alexa MD more accurate and knowledgeable.

Built With

• alexa
• data-mining
• javascript
• python
• voiceflow