Getting Started

This guide walks you through signing up for RINK, uploading your first dataset, training a model, and generating a forecast — all in about five minutes.

1. Create an account

Go to rinkglobal.com and click Get Started in the top-right (or follow the direct link to the registration form).

Fill in:

• First and last name — used for personalised greetings and the avatar initials.
• Email — must be a real address; we send a confirmation link.
• Phone (optional) — saved on your profile, never shared.
• Password — at least 8 characters. The strength meter updates live.

After you submit, check your inbox for a confirmation email. Click the link and you'll be redirected back to RINK and signed in automatically.

Tip. If you don't see the email, check spam, then click Resend confirmation email on the page.

2. Open the workspace

Once signed in, click Workspace in the navbar — or go directly to rinkglobal.com/analytics.

You'll see four KPI cards across the top: Dataset rows, Cadence, RMSE, and MAE. They start as — until you upload data.

3. Upload a dataset

In the left column, find the Upload Dataset card and either:

• Drag a CSV file onto the dotted box, or
• Click the box to open a file picker.

Constraints:

• File type: .csv
• Size: up to 10 MB
• Must contain at least one numeric column

When you click Upload & Train, RINK does three things:

1. Auto-detects a date column. Looks for date, timestamp, time, datetime, ds, or period (any case). Falls back to scanning string columns for parseable dates.
2. Sorts ascending. If your data is reverse-chronological, RINK reorders it so the most recent row is last.
3. Auto-trains a gradient-boosting model and reports RMSE / MAE.

You'll see a green toast confirming success and the chart populates with your time series.

4. Switch columns (optional)

If your CSV has multiple numeric columns, the Model card shows them as emerald pills (or a dropdown if there are more than six). Click a different column and RINK refetches the data, retrains the model, and updates the chart automatically.

The selection is saved per-browser, so reloading the page keeps you on the same column.

5. Generate a forecast

In the Generate Forecast card:

1. Pick a horizon (5, 10, 14, or 30 — labelled in the inferred unit: "10 weeks", "14 days", etc).
2. The Most recent values box auto-fills with the last N values from your dataset. You can edit them or click Use last N to refill.
3. Click Generate <N>-<unit> Forecast.

The forecast appears as:

• A dashed green line on the main chart, starting from "now".
• A shaded green band around it — the confidence interval (±RMSE, widening with horizon).
• A row of forecast tiles showing the predicted date and value for each step.

6. Save and re-forecast as you like

Your data and trained model are saved on the server under your account. As long as you stay signed in, you can:

• Switch columns and re-train at will.
• Generate as many forecasts as you want.
• Reload the page; everything is restored.

Files are deleted automatically when you sign out (or after 4 hours of inactivity). To keep your dataset, simply stay signed in.

What's next?

• Read the Forecasting guide for a deeper look at how recursive prediction and confidence bands work.
• Browse the API Reference if you want to drive RINK programmatically.
• Skim the FAQ for answers to the most common questions.