Offline-First Field Survey Data Collection App

Field surveys happen where the network doesn't: farms, construction sites, remote villages. Paper forms get lost and retyped; most mobile form tools assume connectivity and fail gracelessly without it. The hard problems are offline-shaped: records must queue durably on the device, photos must survive app restarts, uploads must resume rather than restart, and when two people edit the same record the system must surface the conflict instead of picking a winner silently. This project builds an Android app around those problems: a form library with per-form progress, structured data entry with GPS tagging and photo capture, on-device validation, and a WorkManager-driven sync queue with retry, resume and conflict flagging.

1. The supervisor assigns survey forms; the app downloads form definitions and any reference data while online.
2. In the field the surveyor opens a form and works fully offline — entries save to the on-device Room database instantly.
3. Each record captures a GPS fix (with accuracy), timestamp and photos alongside the answers.
4. Validation runs on save: required fields, numeric ranges and formats are checked before the record is queued.
5. WorkManager monitors connectivity and uploads queued records in order when a network appears, with retry and resume.
6. The server acknowledges each upload; records the server has changed meanwhile are flagged as conflicts for review.
7. The supervisor view shows per-form progress, sync states and flagged conflicts across the team.

Tech stack:

• Kotlin (Android app)
• Room (on-device SQLite database)
• WorkManager (background sync scheduling)
• CameraX (photo capture)
• Fused Location Provider (GPS fixes)
• Retrofit with OkHttp (API sync client)
• Django REST backend (reference server)
• Material 3 UI components

• [Form library with progress] Assigned surveys list with field counts, per-form completion percentages and resume-from-last-record.
• [Structured data entry] Text, numeric, single/multi-select, date and GPS field types with skip logic between questions.
• [GPS tagging with accuracy] Every record captures latitude, longitude, accuracy radius and timestamp from the fused location provider.
• [Photo capture] Required/optional photo fields with on-device compression; photos are linked to records and survive restarts.
• [On-device validation] Required fields, ranges and formats are validated before save, so bad data never reaches the server.
• [Sync queue with retry] WorkManager uploads records in order with exponential backoff; interrupted uploads resume instead of restarting.
• [Conflict flagging] If the server holds a newer version of a record, the app flags it for human review rather than overwriting either copy.

• Complete Android app source (Kotlin) with form engine
• Sync engine with queue, retry and conflict handling
• Reference Django REST backend for the viva demo
• Form-definition format documentation and sample surveys
• Supervisor progress views
• Project report PDF (background, offline-first design, sync protocol, methodology)
• PPT presentation for final review
• Viva Q&A preparation document (Room, WorkManager, location APIs, conflict resolution)

• GPS accuracy degrades indoors and under dense cover; the app records the accuracy radius honestly with every fix.
• Very large photo sets can fill device storage; the app warns at configurable thresholds.
• Conflict resolution is manual review — no automatic merge of conflicting records.
• The reference backend is a demo server; production needs the buyer's own hosting.
• Background sync is subject to Android battery-optimization policies on some OEM skins.