One of the biggest problems private duty home-care companies face are issues with their staffing and scheduling. In addition to the caregiver shortage, the staffing coordinators have to match hundreds of caregivers and clients together, and currently, there is no good way to do it. This results in numerous mistakes, missed shifts, wasted time and important information being kept in the staffing coordinator’s head. Our software can reduce the number of staffing coordinators needed in an office by 25-50% which can save owners $45,000-$90,000 a year while maximizing revenue, minimizing scheduling errors, and improving caregiver retention and customer satisfaction.

*The home healthcare industry is one of the fastest growing industries in the country. There are currently 33,000 Home Care Providers in the U.S. alone (Hopkins Medicine). There are 12 million people currently under care and there will be 1.3 million more jobs available by 2020 (IBISWorld, 2017). By 2030, more than twenty percent of U.S. residents are projected to be aged 65 and over, compared with thirteen percent in 2010. In 2050, the number of Americans aged 65 and older is projected to be 88.5 million, more than double its population of 40.2 million in 2010.
70% of people 65 and older will need some kind of long-term care eventually (Forbes). *

The home care industry has historically been very fragmented due to its local and personal nature of the business. Many caregivers want to work close to home with clients that they can build a relationship with. Clients want to work with a local company with caregivers from their community that they can bond and build strong relationships due to the personal nature of the care being received. Our personality matching algorithms are designed to give the local and personal touch, without the local knowledge. Private duty homecare companies no longer need to have local knowledge in order to succeed in a local market.

What it does

We have developed a map-based, easy to use scheduling software that optimizes the match between the caregivers and clients. Our user-friendly system uses algorithms and machine learning to recommend the best possible caregivers for each client. Our algorithms take into account skills, location, wage, background, likes, interests, and personality when recommending matches.

How we built it

We used Docker to containerize our Django application. There are three Docker containers, Postgres for our database, nginx to serve static content and reverse proxy to gunicorn running on our third docker container which is the Django python application itself. We used Bootstrap4 on the front end and the Google Maps JavaScript API to plot markers on a map

Challenges we ran into

Persisting data in a Postgres image on Docker Django's build tools for static files not integrating well with the ephemeral filesystem of Docker containers Adapting a package for user authentication to fit our needs of having users request access to an account rather than signing up and verifying email Modeling the matching criteria in the database and interfacing with the data from ClearCare

Accomplishments that we're proud of

Constructing a RESTful API with the Django REST framework Using our own API to plot markers on a map with the Google Maps JavaScript API Leveraging technology to solve real-world problems

What we learned

Gained experience developing in the Django web framework Locking down an API requiring user authentication to access data Developing with the Google Maps Javascript API Containerizing apps efficiently with Docker

What's next for Smart Staffing

Meet with private duty home care companies to get their feedback and iterate on our software. Reach out to ClearCare, KanTime and other enterprise software companies about partnering with their platforms as a third party package. Unit test our project and integrate with Travis CI. Work out a seamless deployment flow for multiple developers testing out things locally. Optimizing our algorithm to match caregivers and clients more effectively. Implementing a javascript framework like Vue.js.

Share this project: