/ install
Fill in your details below, then download a single file and run it. The installer does everything — clones the repo, sets up Python, creates the database + app user, loads the initial data.
already tried this? restore your values to retry
Form fields aren't auto-saved (passwords stay private even locally). If your install failed — wrong credentials, network issue, a typo — and you want to retry, importing here pre-fills the whole form so you only fix the broken bit instead of re-typing everything.
The data-cycle-installer.py you downloaded last time still has every value baked in (Postgres host/port/db, admin user, app user, MySQL, sFTP, SMB share credentials — passwords too). Drop it here to restore the entire form at once.
If you already deployed once and only have the.env file from the install folder. This restores the parts saved in .env (app user, database, MySQL, sFTP, etc.) but not the admin password or SMB share credentials — those live only in the installer file, never in .env, so you'll re-enter them here.
what you'll be wiring up
before you start, make sure you have:
Where your Postgres instance is running. The installer will connect here to create the app database and user if they don't exist yet.
A user with CREATE DATABASE and CREATE ROLE privileges (usually postgres). Used only during install to create the app DB + app user, then discarded.
⚠️ These credentials are NOT saved to .env — they're only embedded in the installer script you download.
The user and database the pipeline uses every day. The installer creates both if they don't exist and grants the app user full privileges on the app DB.
The school's MySQL database where raw sensor data is streamed. The pipeline reads from here.
Where daily weather forecast CSVs come from. Skip this if you don't have the sFTP credentials — the dashboard will just miss weather-related features.
Comma-separated. Which stations you want forecasts for.
Where live sensor JSON files arrive from the apartments. The installer will mount this as a drive on your machine (Windows only). Leave empty to skip network ingestion.
ℹ️ Windows only. If you're on Linux/Mac, mount manually first (mount.cifs) and fill in the drive letter or mount point below.
Full UNC path to the remote share.
Drive letter on Windows, or a folder path on Linux/Mac.
Where raw sensor JSON files get archived immutably on your machine. The default is a subfolder in the repo.
Can be a relative path inside the project, or absolute.
The installer normally pauses at the end and asks each question interactively. Pre-answer them here for an unattended install — leave on Ask during install to keep the prompts.
Run SMB backfill
Full SMB → bronze → silver run (20-60 min). The dashboards stay empty until this runs at least once.
Open Power BI dashboard
Launches Power BI Desktop with the .pbix already loaded.
Run KNIME predictions
Triggers the motion + consumption ML workflows in headless KNIME (~15 min).
Auto-start watcher on login
Adds the watcher to Windows Startup so the pipeline runs after every reboot.
Start watcher now (new window)
Spawns the watcher in a separate console so the installer can finish. Safe — watcher keeps running on its own.
Launch admin dashboard
Opens the Streamlit admin pane (data freshness, watcher status, BI setup wizard).
🔒 Your credentials are baked into the downloaded file on your computer only. Nothing is uploaded anywhere — the form is fully client-side. Passwords are never saved to browser storage either. The Postgres admincredentials live only in the installer and are never written to .env.