I Made My Onboarding Data-Driven. A Hardcoded List Ate Four of Six Answers.

Lifemaxxing AI is an app that creates a personalized daily task plan based on answers to eight questions during onboarding. The app is built with Flutter and uses JSON files for easier updates. The onboarding process is crucial because the answers provided by the user directly influence the plan generated by the app. Initially, the developer decided to store the questions and options in JSON files to facilitate changes without having to rebuild the app.

However, upon reviewing the code, the developer discovered that four out of six options for question seven had been silently replaced with default options before reaching the plan generator. This issue affected the accuracy of the generated plan, as the default options did not contribute to the task list. The developer traced the issue back to a layer in the code that converts stored user data into the question IDs expected by the config file.

This layer contained guard clauses that prevented certain answers from being assigned to specific questions. The developer added sanity checks to address the problem, ensuring that each question only received valid answers. By implementing these fixes, the developer aimed to improve the accuracy and effectiveness of the app's onboarding process and the resulting task plan.