This is the Number of Reports for the Storm/Hurriance in the area Hattiesburg.
We can see in detail that no more reports of the same type has been made of yet.
911 call can be hectic for the both ends, specially during the time of natural disasters, wild fires and mass shooting. There will be overflow of traffic, as a result they can not help all the people who are in need of help. Hence, we try to ease the burden of the affected people by the use of AI, voice recognition tool. We have tried to make communication with the people and try to solve them go towards the safe place as much as possible. Even we tried to reduce the burden for the 911 dispatcher by reducing the number of redundant reports of the same kind, around same place and time. This will help people from both spectrum.
What it does
It tries to minimize the 911 load and rescue system phase during emergency. It also helps to locate people to safe area when there is no personell available
How we built it
For the backend, we used express.js we purchased a number through Twilio API which was used for making conversation. Then we used dialogflow google cloud based API to understand the intent of people and also understand the entity that was required for front end mapping and listing.
For the frontend, we used react js to create a website which would talk back to the backend using fetch request and bring out all the details and present it to the 911 designated personnel for their ease of use
Challenges we ran into
Training the model for the dialogflow, google cloud based conversational API was difficult and time consuming. Still, many scenarios couldn't be covered due to the time constraint. Also, due to our accent, the voice recognition software/api from twilio couldn't understand use properly to test and debug our product
Accomplishments that we're proud of
We have developed a full stack product in this two days and also incorporated the ML/AI aspect to it, which we had done it before. Being a team of just two, we are proud of how much we have to present .
What's next for 911 Overflow
- We are looking to train the dialogflow model more in diverse topic/conversation and make it as smart and possible.
- Also, we will definitely increase the number of topic that our IVR can converse into and help more people.