While we were brainstorming for ideas, we attempted to identify marginalized communities and understand both the how and why as to why the community that we grow up in can have such a big impact on our lives. As part of the Deloitte challenge, we were asked to conceptualize how the government could connect youth programs and services to Canadians who need them the most. We realized that youth who end up in the foster case system have difficulty accessing other government resources which results in a recurrence of the poverty cycle. Many caseworkers have dozens or even upwards of a hundred of children to monitor, meaning that some that have mental health concerns, can easily slip between the cracks, an occurrence which isn’t rare in the foster care system. 19% of children in the foster care system in Canada suffer from depression, anxiety or withdrawal, signifying a very dire demand to ensure that they are not being neglected or forgotten about (due to the immense amount of other children to account for). For vulnerable populations, such as youth from the ages of 0-18, growing up in an unstable household can have long lasting negative impacts which is what we want to break by creating “matr,” an application that assesses how children are acclimating to a household. This app has been entered for the health tract to emphasize the importance of a positive environment on your mental behaviours.
What it does
Matr as a product is broken down into a suite of tools and features. These include an emotional diary application for the foster children themselves and a database display webapp for the parents and caseworkers, supported by a wide variety of under-the-hood technologies, such as MQTT/Solace, NLP/ Sentiment Analysis and Google Cloud. Upon launching the Android application, the foster child is prompted to enter their credentials (which would be in the form of a code word between the caseworker and the foster child). They would then be asked a series of questions relating to how their week has gone on a wide variety of topics to gauge how welcoming and accommodating their home environment has been. This form of analysis is instrumental in determining how the childrens’ mental health has been impacted by external factors in their lives. Their data is then sent to the front-end web app, which utilizes the “emotional scores” that the NLP component of the Android application determined to show an icon beside each child in a table-setup in the caseworker’s app page, which reflects their emotional states (angry face, to signify extremely dissatisfied, to very happy face, which signifies extremely satisfied). We decided to focus on this specific pain point, as it is very difficult for an individual case worker to manage and remember the emotional states of every single one of their foster children, meaning that a visual and straight-forward dashboard with real-time updates of their children’s states is imperative in ensuring that every child’s needs are accounted for. We used a 5-point Likert scale, having each face represent 2 points on the sentiment analysis scoring system. This was done to minimize confusion when caseworkers/parents are tracking the children’s emotional responses to their current accommodations. The webapp is secure, with basic authentication in the form of a username and password login page, and includes other features such as a persona-based profile of a specific child as the foster parents’ page and an instruction page to speed up the process for first-time users.
How we built it
Challenges we ran into
In terms of the design, we wanted to make sure that this was a tool easily accessible for children of all ages. This constraint is accounted for in the android app with a relatively simple and easy to understand interface. For host families, they are most likely overwhelmed with the process of inviting someone else into their home as their own child which is why we wanted to design a simple interface for the web application as well; allowing them to limit time using this tool but still be effective in providing data that can support whether a match was correctly made or not. The biggest challenge that we faced was that after designing native applications on these two platforms, the integration of APIs and the communication between the web and the android app was quite challenging and time consuming. Only after spending hours with the mentor at Solace and another mentor were we able to successfully integrate the requirements of the Solace challenge which was to build your hack using their message broker technology on these two platforms. The difficulties did not lie in creating these apps separately, but it was conforming to the challenges and integrating specific architecture into a product to use features that we might not have otherwise integrated. Overall, developing that link between both of the applications, as well as the use of Google Cloud for NLP proved to be a challenge.
Accomplishments that we're proud of
This hackathon was one where we were learning in every minute of the 36 hours, as we were all beginners in the technologies that we hacked with (NLP, Android Studio, MQTT/brokers, Google Cloud, HTML/CSS, Sketch). We’re proud of creating a functioning Android application as well as designing a web application that could have a real impact in the world. This is an idea that is feasible and a problem needs to be solved as soon as possible due to the potential long term impact of providing people to grow up in healthy households that have the ability to satisfy their every need. For most of the team, we’re relatively new to the world of hackathons and to create a project that we’re passionate about in less than 36 hours is an accomplishment of itself.
What we learned
I think the most important thing we learn in hackathons is to remember that the design process is iterative and that we’ll be changing our project plans constantly based on the needs and feedback of our customers. For example, we could not decide on a product after brainstorming for a couple hours and once we began to narrow down our choices, the features we wanted to implement were constantly changing based on what our capabilities are and how much we expected to learn in 36 hours.
What's next for Matr
For Matr, we chose to focus on the post matching process because we know that it is at this point that there are the largest number of pain points in terms of obtaining a higher level of care from the foster home/system to a household. We would eventually like to improve the matching process itself through the use of machine learning algorithms. In current times, this matching process is tedious and forces social workers to spend hours mulling over pages of descriptions. By improving the matching process, there will be less children that are sent back to the foster care system due to the increase in likeliness of familial acceptance (stronger matches = children stay in the household longer). Another implementation could be using speech to text, to allow younger children who are unable to write to be able to log their emotional states to the application also. We are aiming to make our interfaces as accessible as possible to everyone.