Your personalised load profile

How to turn it on

There are two ways to feed the predictor — you can use either or both, and the calculator picks the best available source automatically.

Option A: upload your distributor data

1. Upload at least a few months of half-hourly data from your distributor portal — see importing measured electricity data for the how-to. Twelve months is ideal because the model gets to see every season; less still works, it’ll just be a bit less specific outside the months you’ve uploaded.
2. On the calculator’s energy-balance section, leave “Use my measured-data profile” on its default “Auto” setting. The generic load profile dropdown greys out and the daily-load field is ignored — the model already knows your scale.
3. Run the calculation. The chart caption shows the dataset name; the chart itself reflects your household’s usage curve.

Option B: connect your inverter

1. On the Account page, add the inverter (Deye Cloud, GoodWe SEMS, or Fronius Solar.web) under “Connected inverters”.
2. On the calculator, link the inverter connection to your layout via the layout’s inverter dropdown.
3. Leave the “Use my measured-data profile” dropdown on “Auto”. If you haven’t uploaded a CSV, the calculator pulls the last 14 days of load readings from the inverter and trains the profile from those. The chart caption reads “Inverter (Home Deye, 14 days)” (or whatever you named the connection).

When both a CSV upload and an inverter connection exist, the CSV wins (it usually covers more history and a wider range of seasons). Pin to either source explicitly via the dropdown if you want to compare.

What we filter out (and why)

A blind average over your whole upload would be misleading because two situations contaminate the data:

• Holidays. Stretches of consecutive very-low-use days — you’re away, it’s just the fridge running. We spot these automatically and exclude them so they don’t drag your predicted profile down.
• The day you switched on solar (or a battery). Once panels are on the roof, your smart meter only records what’s left over after self-consumption — not what your house actually used. If the predictor sees that pattern in your data, it discards everything from the install date onward and trains on the months before, when the meter still saw the full picture.

A summary of what was excluded comes back with every prediction, so you can sanity-check it against your own memory of when you were away or when the system was installed.

Privacy

Your uploaded data lives in your account row only. It isn’t shared with other users, isn’t pooled into any cross-user model, and isn’t included in the analytics we use to improve the calculator. The predictions we generate from it are sent only to your browser session.

Delete a dataset on the Account page and the model built from it disappears with it. Close your account and the underlying file is purged from our database.

A few things to keep in mind

• Re-upload now and then. Lifestyle changes — new EV, new heat pump, working from home — only show up after you upload more recent data. A fresh quarterly NEM12 is plenty.
• One profile per dataset, not per address. If you moved house mid-year, upload separate NEM12 files for each address rather than one combined one.
• Already on solar? A grid-only CSV from after your solar install only sees what was left over after self-consumption — not your real total usage. We skip post-install months automatically. If you connect your inverter (Option B above), the inverter reports total household load directly and the post-install gap goes away.