Sentinel

Join us to redefine protection and create a safer world for all.

Comment

Inspiration

We are on a quest to create a world where every woman feels safe, supported, and empowered to thrive. Together, we can make a difference and build a future where women's safety is non-negotiable.

What it does

A. Covert Communication for Victims of Abuse:

Sentinel addresses a critical issue faced by women in abusive relationships who struggle to seek help directly due to their partner's constant monitoring of their social media and communication channels.

We are using openAI's generative image to create new images for steganography flowchart

How it helps

Sentinel offers a lifeline to individuals trapped in abusive relationships, providing them with a covert means to reach out for help without fear of retaliation.

Culprit Identification System

Sentinel also draws inspiration from the chilling tales of serial offenders like Ted Bundy. The platform designed to address the challenge of tracking repeated offenders who commit similar crimes across multiple locations by leveraging machine learning models, specifically sklearn and KNN (K-Nearest Neighbors), to analyze and link similar descriptions of suspects across different crime scenes.

For example: Person A with dark brown eyes, tall, around 29ish age, curly long hair is similar to Person B with black eyes, medium, 30 years age, long black hair. Here we use ML to find the similarity in the text

Additionally, Sentinel provides intuitive visualization tools to map out the locations of these incidents, offering law enforcement agencies a comprehensive view of the offender's activities. Flowchart

How it helps

Sentinel revolutionizes the way law enforcement agencies tackle crimes by providing them with powerful data-driven insights. By consolidating information from multiple crime scenes and identifying commonalities in suspect descriptions, our platform enables proactive measures to apprehend offenders more effectively. Moreover, Sentinel's visualization features offer a holistic understanding of the geographical distribution of criminal activities, facilitating strategic resource allocation and crime prevention efforts.

How we built it

  • Database: Convex db,
  • Frontend: React js, tailwind, bootstrap
  • Backend: Flask, Swagger, Serve
  • ML: OpenAI, KNN, SKlearn,
  • Dynamic map generation: folium
  • Steganography: Python, PIL

Where is Convex db used

  • Clerk auth and convex is used to authenticate the user into the system

Data storage for real time updates:

  • Use for creating a centralized db for crime reporting with all CRUD features
  • Storing decrypted SOS messages

A. Covert Communication for Victims of Abuse:

  • The convex table - messages get updated everytime there is a telegram post with an image and hastag #staywoke

B. Culprit Identification System

  • The users can report a new culprit into the table called culprit. We could get the individual values like eye color, hair, age, other details from the user and condense it into a single message & add it to the db
  • Then they can see the detailed view and search for the report
  • It also supports the deletion. - Hence all the CRUD functionality is supported here

Screenshots of sample databases used

culprit

encoded

Flask server flask

Challenges we ran into

  • A lot of issues while connecting convex db in my windows system. Had to spend a lot of time troubleshooting
  • We were using Python (flask) for certain steganography and ML models. But we were getting some issues like deploying the backend. Hence we were not able to deploy the entire code into production. We tried to lot to deploy the entire stack, but were getting a lot of blockers. So please excuse us for not providing the deployed URL
  • Shortening the demo pitch to exactly 3mins also was a challenge

How run the code on local

  • Install the required packages in react and python

For python

pip install -r requirements.txt

For react

cd Convex
npm install

  • Generate a new openAI key and export it (It is used to img generation)

    export OPENAI_API_KEY=<Key>

  • Update the .env file in apps/web with the following values

    NEXT_PUBLIC_CONVEX_URL=
NEXT_PUBLIC_CLERK_PUBLISHABLE_KEY=
CLERK_SECRET_KEY=

  • To display the map we require serve package. For that do the following

    npm install -g serve

    In the parent directory, execute the following

    serve

  • Generate a new token forthe telegram bot and add it to a channel. The token needs to be exported

    export TELEGRAM_TOKEN=<TOKEN>

    Then open a new terminal and run the dvtelegram.py function

  • Open another terminal and start the react convex server and specify which db we want to use (either new one or an existing one)

    cd Convex
npx convex dev

Run the following command to start the server. This would open the app at http://localhost:5000

npm run dev
  • Open a new terminal to start the flask server python main.py This would start a new flask swagger app. To see the URLs, go to http://127.0.0.1:5000/apidocs/

Built With

Share this project:

Updates