flowchart LR
    subgraph MOBILE["Mobile app (Flutter) - patient"]
        direction TB
        UI["Presentation<br/>go_router, flutter_bloc Cubit, fl_chart"]
        DOM["Domain<br/>entities, repository interfaces, use cases"]
        DATA["Data<br/>repository impl, DTOs, supabase_flutter"]
        HIVE[("hive_ce<br/>local results + outbox<br/>flutter_secure_storage: session")]
        subgraph DSPM["DSP module (isolate, on-device)"]
            direction TB
            SENS["sensors_plus<br/>accelerometer ~100 Hz"]
            PIPE["QC gate, resample 50 Hz, band-pass 1-20 Hz<br/>Welch PSD, features, rule-based decision"]
            SENS --> PIPE
        end
        UI --> DOM --> DATA
        DATA --> HIVE
        DOM --> SENS
        PIPE --> DOM
    end

    subgraph WEB["Web dashboard (React Router) - neurologist"]
        direction TB
        WUI["Routes / loaders<br/>patient list, trend charts, medication timeline"]
        WCL["supabase-js client<br/>email + password session"]
        WUI --> WCL
    end

    subgraph SUPA["Supabase backend"]
        direction TB
        AUTH["Auth<br/>anonymous sign-in (patient)<br/>email + password (doctor)<br/>JWT: role authenticated"]
        API["PostgREST + RPC<br/>connect_to_doctor, disconnect,<br/>create_measurement (idempotent on id),<br/>list_my_measurements,<br/>list_patient_measurements"]
        subgraph PG["Postgres + Row Level Security"]
            direction TB
            T1[("profiles<br/>id, role, display_name, connection_code")]
            T2[("patient_doctor_links<br/>patient_id, doctor_id, linked_at, revoked_at")]
            T3[("measurements<br/>status, features jsonb, algo_version,<br/>calibrated, last_dose_at, device jsonb")]
            T4[("medication_schedules / medication_logs<br/>contract pending")]
        end
        AUTH --> API
        API --> PG
    end

    DATA --> AUTH
    DATA --> API
    WCL --> AUTH
    WCL --> API
    API -.-> WCL

    classDef store fill:#e8f0fe,stroke:#2941a3,color:#141a2e
    classDef pending fill:#fff5c2,stroke:#735c00,color:#735c00
    class HIVE,T1,T2,T3 store
    class T4 pending

1. A patient opens the mobile app and runs a short tremor check with the phone resting on a supported hand. No sign-up is needed to start, and the app works offline.
2. An on-device signal processing module filters the accelerometer data, analyzes its frequency spectrum, and applies a transparent, rule-based decision, so every result can be explained and thresholds can be tuned without any black-box model. Raw sensor signals never leave the phone.
3. Once the patient links to their neurologist with a connection code, the extracted features and result sync to the backend, and the medication schedule set by the doctor on the web dashboard appears as reminders in the app. Dose logging is captured on the phone too.
4. The dashboard presents each patient's tremor trend alongside medication timing, closing the loop between screening, treatment, and monitoring, so neurologists review objective patterns between clinic visits instead of relying on memory alone.

The NuroPulse web dashboard can be accessed here. The mobile app (for now, only the Android APK build is released) can be installed from here, download the latest .apk file to your Android phone, and install it.

To ease the process of testing, we have seeded a couple of users in the backend:

Built by nama timnya apa, a 4-person team for JOINTS Hackathon 2026 from the Faculty of Computer Science, Universitas Indonesia 💾

	Name	Institution	Role
	Risyad Athaya Muhammad	Information Systems '25 @ Universitas Indonesia	Team Leader & Backend Engineer
	Mohammad Zidane Kurnianto	Information Systems '25 @ Universitas Indonesia	Mobile Engineer
	Rayna Kayla Rayvanka	Computer Science '25 @ Universitas Indonesia	Business Lead
	Tania Ju	Information Systems '25 @ Universitas Indonesia	Frontend Engineer

Signal processing

• Welch, P. D. (1967). The use of fast Fourier transform for the estimation of power spectra: a method based on time averaging over short, modified periodograms. IEEE Transactions on Audio and Electroacoustics, 15(2), 70–73.
• SciPy documentation. scipy.signal.welch, butter, sosfiltfilt. https://docs.scipy.org/doc/scipy/reference/signal.html

Tremor characterization

• Deuschl, G., Bain, P., & Brin, M. (1998). Consensus statement of the Movement Disorder Society on Tremor. Movement Disorders, 13(Suppl 3), 2–23.
• Bhatia, K. P., et al. (2018). Consensus statement on the classification of tremors, from the Task Force on Tremor of the International Parkinson and Movement Disorder Society. Movement Disorders, 33(1), 75–87.

Accelerometer-based Parkinson's data

• Varghese, J., et al. (2024). Machine learning in the Parkinson's disease smartwatch (PADS) dataset. npj Parkinson's Disease, 10, 9. https://www.nature.com/articles/s41531-023-00625-7

• ANTARA (2 May 2026). PERDOSNI siap isi kebutuhan dokter saraf RS di daerah. https://www.antaranews.com/berita/5551344/perdosni-siap-isi-kebutuhan-dokter-sarafrs-di-daerah

• sensors_plus (Flutter accelerometer package). https://pub.dev/packages/sensors_plus