SafeStreets

Logo

Inspiration

Being a student in high school in after-school clubs, we often have to return home late, even during the dark. In our relatively safe community, we're still frightened by the possibilities of danger surrounding us. For example, just a few weeks ago, one of our teammate's aunt was robbed of her gold necklace. We wanted to find a way to improve our neighborhood's safety and our community's quality of life. This is an idea that all of us strongly believe in, as we are passionate for making the community we live and thrive in as safe as possible.

What it does

Our groundbreaking new app is designed to enhance community safety. Our app utilizes the power of Natural Language Processing and word embeddings to scrape social media for information on which streets in a neighborhood are unsafe and presents this data on a map for easy access.

Our app also empowers users to share their personal experiences and insights about each street, creating a dynamic and continually updated resource for all users, resulting in a number rating for 1-10 ranking the safety of each street. Additionally, when deciding to walk on a street, users can see "tags" representing the general reasons why each user deemed a street unsafe. These tags are autogenerated using GloVe embeddings, creating a seamless experience for users.

Our app goes the extra mile to ensure that you are protected. It will plot the safest path from one point to another, taking into account the unsafe streets that you should avoid, and directing you along the safest possible route, while still ensuring that the path is as short as possible.

How we built it

Through the use of multiple APIs and Artificial intelligence applications, we were able to form a robust safest route-finding app. More specifically, we communicated with the Google Maps API to load a dynamic map service to our user, which would allow for real-time location tracking. In addition, we utilized the geocoder modules to efficiently locate the geographical location of the streets and roads following our route. Also, through the use of a finetuned sentiment analysis Natural Language Processing model called BERT, we were able to accurately predict regions with the highest risk of crime in the given user's area. As for the incident reports we receive from the user, they are parsed and summarized through an innovative take on the use of GloVe word embedding. While using methods such as cosine similarity and L2 decay distance metrics, we were able to very simply and effectively extract relevant categorical tags to each of the user's inputs. In combination, each of the subsections communicated through a robust REST API server implemented through Flask, which stored all of the data in a SQL database, allowing for a scalable solution.

Challenges we ran into

During the hackathon, our team encountered several difficult challenges that tested our problem-solving abilities. One of the most significant challenges we faced was working with the Google Maps API. Despite having some prior experience with API integration, we found ourselves struggling to implement the advanced features we required for our application. We spent a significant amount of time searching through the API documentation and experimenting with various code configurations until we were able to achieve the desired functionality.

Another challenge we faced was designing a visually appealing and user-friendly interface for our app. While we possess technical skills, designing an effective UI was not an area of expertise for our team. We had to revise and fine-tune our design multiple times, seeking feedback from others and incorporating different suggestions until we were satisfied with the final result.

Accomplishments that we're proud of

As a team, we are immensely proud of the advanced artificial intelligence system that we developed in conjunction with our backend. We believe that it is a true testament to our technical expertise and problem-solving abilities. Our AI system is a sophisticated and innovative solution that utilizes cutting-edge technology to provide users with valuable information on the safety of their neighborhoods.

In addition to our technical accomplishments, we are also extremely proud of the teamwork that we demonstrated throughout the hackathon. As a team, we worked collaboratively and communicated effectively to overcome obstacles and achieve our goals.

What we learned

Throughout the event, we were able to gain experience in areas that none of us had prior knowledge of. For instance, we lacked expertise in user interface design, but by working together and learning from one another, we were able to identify what works visually and what doesn't. We were also able to enhance our mobile application development skills significantly by dedicating a considerable amount of time to the project, and we learned how to seamlessly integrate our server with our mobile app using JSON requests for a REST API.

What's next for SafeStreets

While we are proud of the progress that we made during the hackathon, we recognize that there is still significant potential to improve our app and make a meaningful difference in the community we live in. As we continue to develop our app, we will be focusing on refining our server implementation, improving the scalability of our system, and ensuring that we have access to the necessary resources to support the ongoing operation of the app.

Additionally, we recognize that the UI design of our app can be improved. While we are proud of the progress we made in this area during the hackathon, we will continue to work on refining the design to make it more user-friendly and visually appealing.

Finally, we are committed to enhancing the functionality and effectiveness of our artificial intelligence component. We recognize that with more computing power and more advanced technology, we can make even more accurate and valuable predictions about neighborhood safety.

Ultimately, our goal is to create an app that truly makes a difference in the community we live in. We are committed to continuing to develop our app and refine its features until it becomes a powerful tool for promoting safety and security in our neighborhood.