Pawsitive Care

Inspiration

The inspiration behind Pawsitive Care was a shared passion for improving the well-being of pets and making pet care services more personalized and effective. Recognizing the growing demand for innovative solutions in the pet care industry, our team set out to create a platform that not only enhances the experience for pet owners but also helps pet care service providers better target their audience.

What it does

Pawsitive Care utilizes cutting-edge AI and segmentation techniques to offer a comprehensive pet caring service. It combines predictive analytics with recommendations to warn user about potential health issues in pets at an early stage, saving both pet owners and service providers time and money. The platform focuses on pet boarding and veterinary services, aiming to match the right services with the right user based on their unique needs and characteristics. In our project, we rely on historical data to predict users' future actions. For instance, we are working to forecast when a person will need pet boarding.

Methodology

Segmentation forms the cornerstone of Pawsitive Care's approach, facilitating the classification of users into distinct categories based on their preferences, habits, and pet-related characteristics. The subsequent analysis of user-generated data, obtained through lead generation forms, serves as the foundation for personalized service recommendations. The platform harnesses machine learning models, trained on simulated data, to predict optimal service timings and cater to the individualized needs of users.

The predictive capabilities of Pawsitive Care extend to both veterinary and pet boarding services. Real-time recommendations for veterinary services ensure timely intervention and care, contributing to enhanced pet health outcomes. In the context of pet boarding, the platform strategically utilizes AI to comprehend user habits and lifestyle, subsequently tailoring promotions to align with individual preferences. For instance, targeting users with a proclivity for travel and a lack of suitable pet care arrangements allows for precise marketing strategies. Historical data pertaining to boarding patterns further refines the platform's promotional endeavors, offering predictive insights into future boarding needs. The variables we consider include their past usage of boarding services, travel preferences, upcoming holidays, and demographic data. It's important to note that the historical data used to train our model, which includes information about their travel and boarding service usage, is artificial mock data.

How we built it

We built Pawsitive Care on a robust architecture that leverages various technologies. The system collects demographic and behavioral data through the Segment platform and supplements it with geographic and psychographic data from external sources (mock data). All this information is stored in Snowflake, ensuring a secure and scalable data storage solution. The AutoML feature of Databricks is employed to train and identify the best machine learning model using historical and external data (mock data). After using ML to predict user needs and behavior, we write results back to snowflake. For sending emails we use Loop which connected to a Reverse ETL Snowflake warehouse.

Challenges we ran into

During the development process, one of the major challenges we faced was integrating real-time data analysis. Initially, we attempted to use Databricks Delta Live Table as a source, only to realize that it wasn't ready for our specific needs. This setback prompted us to reassess our approach and find alternative solutions to ensure seamless insights as fast as possible.

Accomplishments that we're proud of

Despite the challenges, we successfully implemented a system that predicts user behavior and needs based on historical data. The system's ability to provide personalized recommendations, using the right content for the right user, has enhanced the overall user experience. The integration of predictive analytics, recommendation systems, and machine learning not only enhances the efficiency of service provision but also contributes to the overall well-being of pets.

What we learned

The journey of building Pawsitive Care taught us valuable lessons about adaptability and resilience in the face of technological challenges. We gained a deeper understanding of the importance of choosing the right technologies for real-time data analysis and the significance of continuous learning in the rapidly evolving field of AI and machine learning.

What's next for Pawsitive Care

Through partnerships with local animal shelters, veterinary clinics, and pet-focused organizations, our platform may become revolutionizing marketing innovation. Looking ahead, our focus is on further development and refinement. We plan to implement a real-time recommendation system using a stack that involves Segment, AWS Firehose Data Streams, Databricks notebook, Databricks DLT, and Loop. Additionally, we aim to automate the training of ML models each week with fresh data using Databricks AutoML scheduler. Future enhancements include the incorporation of additional ML models to predict various pet diseases, making Pawsitive Care an even more comprehensive and proactive solution for pet owners and service providers alike.

Built With

• amazon-web-services
• css
• databricks
• html
• javascript
• loop
• particle
• python
• segment
• snowflake