The domain the Nessie API is hosted on is called "". Through Lapras, we've created a unique experience that allows users to communicate and obtain vital information about their personal finance, which is both more convenient and engaging that simply accessing information from mundane tables. We named the project Lapras because of our admiration for the Pokemon Lapras, and also because of its similarities to Nessie.

What it does

Lapras is a web robot that utilizes Capital One's Nessie Hackathon API, and emulates interactions that can be made with personal bankers. Lapras can make queries to filter transaction activity between a set time frame, and make predictions based on spending patterns.

How we built it

Lapras is powered by PHP. The frontend was designed in HTML and CSS, which was then embedded into the PHP backend. Transitions often utilize JavaScript. All graphic elements, including icons and sprites have been designed on Adobe Photoshop, Illustrator and After Effects.

Challenges we ran into

Building a web robot makes us enter the broad realm of Natural Language Processing. Making a script understand what the user desires from just a string is a difficult parameter to base your script upon. Nonetheless, we used keyword searches and identified ways to implement features that we recognized were essential to making this finance bot.

Accomplishments that we're proud of

We're proud of how we have created this tool without any prior experience in building bots or related scripts. In addition, we're also happy with how the user interface has come along, and seamlessly integrated with the backend.

What we learned

We were not sure about processing times related to making requests, but the well-documented Nessie API helped us understand how requests could be made to it and what parameters could be evaluated. We also looked at different ways of implementing primitive language processing, and we're definitely going to explore this in the future.

What's next for Lapras

As we've mentioned earlier, Lapras needs to continue expanding it's memory. Our goal is to make it understand more sentence structures, and develop scripts that automate this learning process for it. In addition, we would also want to include more predictions and projection based queries. Finally, we also need Lapras to be able to make transactions of its own, and in the future, we hope to utilize more of the API to enable this.

Share this project: