With the API's provided and the challenges set we endeavored to create a project to that tackled the problem of improving adult social care by using financial data for good.
What it does
It analyses the financial transactions of a patient. Using this data it can spot anomalous behavior and alert a social worker to this. This can help social workers better allocate there time to patients that are doing activities that are out of the ordinary. as a change in behavior may indicate a degradation or potential improvement in their condition worth a checkup.
How we built it
Using a Node.js back end we interface with the Capital-One Hackerthon API. Every day at 9am we analyse the expenditure of patients and identify 'unusual' behavior through statistical analysis. Every user of concern should then be mapped to their social worker. Then their social worker is sent a text, alerting them they may want to check in with those patients. If a social worker wants to know more information about one of the patients they've been alerted to, they can reply to the text which runs a twilio webhook and they will get a text back with more information. We have built a express webserver to give a webportal to social worker allocation.
Challenges we ran into
Coordinating the sleep deprived team. Using many new API's there was a lot to learn.
Accomplishments that I'm proud of
All of the API interfacing works. Making the base of a system that could be genuinely useful in the real world.
What I learned
Lots about REST API's
What's next for Plutus
Fleshing out the functionality. Training AI models on real data to out preform our current statistical analysis