CrowdFlow

Inspiration

We were inspired to bring to the advances made in data science and machine learning to event avenues such as cruise ships, where understanding the movement patterns of the crowd could help event planners make better decisions to improve the guests experience.

What it does

*Uses the existing security cameras on board the ship to leverage computer vision and count how many guests are in each area of the ship at each time. *The collected data is stored in a database that represents that particular trip. Such databases are collected from multiple trips made by multiple ships and used to train machine learning models that can predictions on crowd movements in future trips. *Despite the powerful computation and connectivity behind the scenes, the user interface for administrators is very simple. Allowing for easy access to valuable insights.

How I built it

*Database - We used google cloud SQL to host the database. We decided to use a relational database so that the machine learning model could interpret them more efficiently. *Google Cloud - we used google cloud services throughout the project to help create a full stack solution. On the backend we have a Google SQL database with a mysql engine that connects to the frontend app using cloud functions to ensure a more secure solution. The frontend is also hosted on google cloud using App Engine. Thus, the full stack is built solely on Google Cloud.

Challenges I ran into

The biggest challenge we faced was how new we were to using these technologies. For all of us, this was not only our first web app, but our first time making a full-stack solution with a specific client in mind. The pressure from not being on par with the clients needs pushed us to do our best and overcome the numerous bugs and sticky situations we found ourselves in.

Accomplishments that I'm proud of

As a team, we are most proud of venturing into the unknown and taking a risk to create a solution that mattered to a client. About 75% of the project, we learned how to do during the competition. We were all dedicated as a team, held our own weight, and trusted that everyone would be able to persevere and get their part of the project done. That team bond is something we all appreciated as we only met at this event and had no real prior knowledge of each other's abilities.

What I learned

Technology-wise, we learned React.js, mysql, SQL, node.js, Google-cloud, and Figma prototyping. But metaphysically, we learned the importance of commitment, determination, and caffeine. These are lessons we will continue to learn and develop as we continue to attend more MLH events and create new memories.

What's next for CrowdFlow

Ideally, we would like to expand this solution to other venues past cruise lines such as concerts, conventions, university events, and shopping. We would also like to investigate other forms of data collected to incorporate into our model for crowd flow prediction. There are a lot of applications for this product and it can be easily tailored to any client.

Built With

• app-engine
• google-cloud
• google-sql
• mysql
• node.js
• react