We are the students and professionals who worry about inconsistencies between what was planned and what actually happens. We want to build a smart system that dynamically rearranges your todo-list so that it reflects actual working conditions, instantly changing tasks priorities and your current state.
What it does
This is an application that turns your phone into a second screen displaying your current task and shortlist of the next tasks. When you finish the task or want to make a break, it requests for your feedback to make the best personalised list for you, where tasks of different types are mixed in order to prevent you from getting bored with monotonous work and be as effective as you can at the same time.
How we built it
This application uses Deveo as a source of tasks, MongoDB, Java for building Android app and Python for business logic. The backend uses machine learning methods to build an optimal task rearrangement algorithm. The machine learning part consists of two stages: firstly, all the tasks are getting marked with labels according to their description in natural language. After that, for each user a best sequence of tasks is chosen from a pool of tasks waiting in a todo-list considering tasks descriptions in terms of labels, task priorities and user's feedback on each completed task.
Challenges we ran into
Different vision of problem and ways of solution suggested by each member of the team that caused some difficulties on the stage of integration all the parts of application; common problems with communication in English as our team was international; almost loosing one member of the team because of sudden good weather on Saturday.
Accomplishments that we're proud of
Finding a solution that satisfied all of after series of arguments and finally creating a working multicomponent prototype.
What we learned
Using Deveo API, applying instruments of machine learning to real problem(not training one that was specially constructed in order to practice a particular ML method).
What's next for SmartPlan.tech
Change priority-feedback metrics from hardcoded to adaptive and improve global tasks classifier after collecting more feedback, making application avialable for iOS and Windows Phone.