sonnyAngels

Inspiration

Our inspiration for the SonnyAngels project stemmed from a desire to revolutionize the world of online advertising. We were motivated by the untapped potential for optimizing ad-moderator allocation to significantly increase revenue. This quest was fueled by our belief in the power of data-driven decision-making to transform advertising operations and drive tangible results.

What it does

SonnyAngels is a sophisticated system designed to intelligently allocate advertisements to moderators. Its core function is to optimise the utilisation of moderators, leveraging their unique strengths while considering essential factors like real productivity and accuracy. The system employs a simulated annealing algorithm to explore and fine-tune allocation scenarios, ultimately achieving a more efficient and effective ad-moderator assignment.

How we built it

The journey to build SonnyAngels was a multifaceted endeavor. We started by cleaning and preparing our data to ensure accuracy and consistency. Next, we developed scoring models that factored in various criteria, such as real productivity and accuracy. The implementation of the simulated annealing algorithm was a pivotal step, allowing us to intelligently reassign advertisements while gradually reducing temperature to fine-tune allocation decisions. This was a complex undertaking that required meticulous design and rigorous testing to ensure optimal results.

Challenges we ran into

Building SonnyAngels came with its fair share of challenges. Fine-tuning the algorithm to strike the right balance between allocation and temperature control proved to be a delicate task. We encountered debugging challenges and had to optimize parameters to ensure seamless data integration. Overcoming these hurdles demanded patience, collaboration, and creative problem-solving.

Accomplishments that we're proud of

Our proudest accomplishment is the creation of a system that promises to elevate utilization by over 10% and increase revenue significantly. We're proud of the innovative approach we've taken to transform ad-moderator allocation and enhance the efficiency of online advertising operations. This project marks a significant milestone in our journey towards harnessing the power of data for tangible business outcomes.

What we learned

Throughout this project, we gained invaluable insights into data cleaning, scoring models, and optimisation algorithms. We learned the art of balancing various factors to achieve a harmonious allocation. Additionally, we honed our debugging and problem-solving skills, reinforcing the importance of persistence in the face of challenges.

DISCLAIMER

Unfortunately, the free plan on Render doesn't have enough memory to process the algorithm in our web application. Hence, we recommend cloning the demo's repository on GitHub (trwstin/AMMO) and running it locally.