JET-LAGGED PLANNER
An automated solution for making a bunch of appointments across different time zones. Built for a wide range of industries including Manufacturing, Aviation, and Construction, this smart planner tackles bulk scheduling with timezone conversion, conflict prediction, data visualisation, and AI-aided rescheduling customised to your own historical data to keep you away from a 3 AM site visit in a heatwave!
MAJOR CUSTOMERS
Worldwide
INDUSTRY
Manufacturing, Dispatching, Hospitality, Facilities Management, Aviation, Construction, etc.
Problem Area
The goal of this project was to design a web app that optimises and streamlines the process of bulk scheduling inspections for large organisations operating across various time zones. The design needed to reduce human error, improve efficiency, and predict conflict resolution strategies by incorporating factors like weather conditions, historical data, operational hours, and personnel availability.
The project stems from one of the customer pain points that I observed during my recent product design internship as the current workflow of inspection scheduling for the Safety Culture platform does not allow multi-national organisations such as Amazon delivery services to assign timezones for a bulk of sites, which can be a tedious task and is prone to human errors. I completed three iterations of work flow wire framing but none of the proposed ideas are ideal for bulk scheduling and increases scheduling efficiency significantly.
Technical Debrief
The model is trained based on randomized data that includes entries like Site Name, Location, Time Zone, Inspection Type, Preferred Time Slot, Inspector Availability, and Weather Condition. And this is to foreshadow the possibility of allowing users to upload their own CSV, and the model will train dynamically on this data (this process is demonstrated through a basic Stream lit web-app).
Utilized K-Means Clustering to group inspection sites based on:
Weather Conditions
Inspection Type
Result: Simplifies Scheduling by grouping inspections based on type and weather which helps identify patterns and categorises tasks efficiently.
The code uses a Random Forest Regression model to predict optimal scheduling times based on K-Means Clustering, Weather Condition, Preferred Time Start, and Inspector Time Start.
Result: Reduces the risk of overlapping tasks by considering key factors like weather and inspector availability.
