Memory loss regarding close people and relatives is one of the most social and emotionally impacting aspect of conditions like dementia or alzheimers, not only for the person, but also for the family and caregivers.(https://www.agingcare.com/articles/alzheimers-parent-does-not-recognize-family-144197.htm). This family angle seems, however, not to be addessed as a central focus for the majority of the current tools we could find. Therefore, this is an app to help users with these kind of conditions and their relatives help them retain or reviewing certain memories and facts about the people who are closest, in order to facilitate coping with this situation and retain them for as long as possible.
Navigation through family members, with picture, name and kind of relationship; Detailed view for a specific family member; Aging picture for some family members, facilitating the connection between the "image the person remembers" and the current aspect of the person;
Family quiz - Gamified memory training with focus on family members;
- Adding family tree representation
- Improving quiz adaptive algorithm
- Adding a trigger mechanism so it must not rely on action initiated by the user
We used React to create reusable components and Redux to manage and control the flow of state. We used aphrodite for CSS, allowing us to package the styles up in the components themselves as JavaScript objects.
In order to improve the effectiveness of the quiz section, we decided to give priority to questions that had been asked furthest in the past and also those with the most incorrect answers. To achieve this, we have implemented a custom algorithm that will determine which question to present the user, based on the user's previous answers.
The code related to this can be found here and here.
We are using the QuickGraph library to represent and store our data related to the user's family. This allows us to use a bidirectional graph to represent people, and the relationships between them.
With no previous experience using a graph data structure, this was a bit of a risk, but we believe it has paid off, allowing an intuitive representation of a family tree, and would serve us well in future features.
The code we use for this can be found here.