Lighthouse

Inspiration

Leading a team is not easy. When managing larger teams it’s hard to track current up to date progress of each individual and to estimate the probable completion dates of tasks and projects against set due dates. Monday is a great tool for organising projects and teams, and gives you command over a lot of processes. We wanted to build this application layer on top of monday to give team leaders, and also team members, a data-led way to anticipate project and task delays and make sense out of the might that monday gives. We wanted to provide a way to support team leaders in having foresight about when a project may fall behind, in deciding when to reassign a task, and track how on-schedule any task/project is.

How the app works

The app acts as a feedback vector for team leaders and teams to see if users and projects are on schedule. It supports team members to stay on track by anticipating delays through forecasting on past metrics like status reports from individuals.

The app takes the past / up to date track record of individual team members and calculates their expected progress for a given time period ahead and will predict if there will be any delays. This is accomplished by using our machine learning model developed using tensorflow.

On the basis of that prediction, it will suggest the most suitable team members to support a task / that team member who is experiencing a delay, or help decide if a task should be re-assigned. We use a team overview dashboard to give team leaders and teams a way to see who has extra capacity to take on new tasks or help another team member to finish on time.

Challenges we faced

Creating a model that can give accurate output with relatively little data was an initial challenge. To solve this we created training datasets in monday boards and also created additional datasets. Working with the monday api and graphql were also new concepts for all of the team members.

What's next for Lighthouse

In the next iteration we would like to expand capabilities such as create a way for team leaders to provide direct feedback to team members at the point of the predictive dashboards. We also would like to support multiple board and project integration.

The app should in future also be able to suggest the most suitable team members to support a fellow team member based on matching skills. We are half way towards this goal in that we have partially implemented this in our codebase and we next need to fully integrate that into the app.