Petsitting and user analytics.

A real-life solution to day-to-day problems of requiring pet sitters and ensuring they are trusted. This solution allows AI checks for similar photos, fraudulent checks and offensive content reporting

Comment

Inspiration

Our pets need the utmost care when we go on vacation or have a long business schedule that requires us to be out of the house for some days. If relatives or neighbors are unavailable and pet homes have cages that many pets are not used to, we need pet sitters to care for the pets in a good environment.

What it does

We provide a medium for pet owners and sitters to communicate and provide pet-sitting services. Our website messaging will help pet owners and sitters communicate. To ensure trust, all users should provide images of available pet foods and proof of pets being taken care of before an arrangement.

The website is designed to enhance the safety and reliability of a pet-sitting service platform by incorporating offensive content detection, image similarity analysis, and data reporting. The backend integrates Azure Open AI, Azure Event Hubs, and Microsoft Fabric to automate the analysis of user messages and images and generate visual reports based on detected issues. The reports are accessible via Microsoft Fabric, which provides reports of potentially harmful content and similarity analysis results.

How we built it

We built it using :

  • ReactJs for the Frontend part of the website. This ensures reliable user interaction
  • NodeJs for the Backend part of the website. This provides routes for communication
  • MongoDB for database
  • Azure portal for Function Apps, Event Hubs metrics, and Azure AI services
  • Microsoft Fabric for analytics

Key Features

  • OpenAI Offensive Content Check: Uses OpenAI's GPT model to identify offensive language in user messages.
  • Image Similarity Analysis using Azure cognitive services: Detects the similarity between uploaded images to check for potential duplicates or related images.
  • Event Hub Integration: Sends offensive content analysis and image similarity results to Azure Event Hubs. Within the Azure portal, Event Hub’s Monitor section can show event flow, providing insights into data volume and latency.
  • Microsoft Fabric Reporting: Streams data to Microsoft Fabric for visualization and reporting, allowing administrators to monitor flagged content and similarity results.

Installation

  • Clone the repository.
  • Install dependencies using:

cd frontend
npm install

cd backend
npm install

Challenges we ran into

  • Microsoft Fabric Trial has no Copilots therefore it could not be incorporated.
  • Render hosting takes a lot of time to load the website because the free instance takes time
  • Azure Open AI standard resource cannot process many requests in one minute so we have to wait for the similarity results for a maximum of 21 minutes if all the similarities are between 40-99%
  • Messaging feature was hard- coded and it was a bit difficult to integrate the logic

Accomplishments that we're proud of

  • I was able to build and host the website.
  • The key features are well integrated.
  • The messaging function can work and the website is functional
  • I was able to complete the challenge

What we learned

  • I learned how to use Microsoft Fabric streaming datasets
  • I can integrate Azure portal resources to work well with Node JS and Microsoft Fabric
  • I learned to use Azure Open AI for sentiment analysis and Cognitive services to calculate message similarity.
  • I learned how to create a chat function in Node JS

What's next for Pet-sitting and user analytics.

  • We need to set up office locations for escrow services between users to ensure trust
  • We need to upgrade our resources to ensure they receive more requests
  • Users can enable other users to view images uploaded from their preferences.

Built With

Share this project:

Updates