Mini Project: A Home Energy Audit, PV, and Your First Savings Plan
Mini Project: A Home Energy Audit, PV, and Your First Savings Plan
You've worked through eight lessons before this one: from "why measure at all" to a dashboard that shows you, every day, where the energy is coming from and what you can do about it. Now it's time to pull all of that together into one practical home audit and turn it into a concrete plan.
This lesson is the module's closing mini project. We're not adding new integrations or another HACS card. You're looking at what you already have, and answering three questions: what do I actually know about my home, what can I improve, and what's genuinely worth doing first.
Plan on about 90 to 120 minutes. It's best to do this on a quiet day when you have the Energy Dashboard, your own dashboard, and a notebook close at hand. By the end you'll have a home energy report: one conclusion, and one plan of action for the days ahead.
The rule for this lesson
By the end of the module, you shouldn't just have data.
You should be able to pull one specific conclusion and one specific plan of action out of it.
What you're gathering for the mini project
This mini project isn't a YAML exam. It's an inventory of what you've built across the module. You don't need to have everything. What matters is knowing what's missing, and what that means.
The Energy Dashboard: import, export, PV, water, gas, and devices added to the built-in view.
Your own energy dashboard: the view from Lesson 8, even if it's still simple.
Daily sensors and cost: utility_meter, a cost template sensor, an input_number holding your energy price.
A device measurement: at least one outlet or Shelly Plug from Lesson 4.
PV and the balance: production, import, export, self-consumption, self-sufficiency, if you have solar.
Automations: at least one from Lesson 4 or Lesson 7, even if it's only a notification.
Dynamic pricing and your rate plan: optional, if you have a pricing integration and know your utility agreement.
You don't need the full house
If you don't have solar, whole-home measurement, or dynamic pricing, you can still run a meaningful audit based on the Energy Dashboard, one device, and your own observations. The report will simply describe what you have, and point to what's worth measuring next.
Data from the Energy Dashboard
Start with the built-in Energy Dashboard from Lesson 2. It's your archive and your point of reference. Open the daily, weekly, and monthly views. Write down actual numbers, not just an impression of "a lot" or "not much."
| What to check | Where on the dashboard | Question for the report |
|---|---|---|
| Electricity use | Grid / home consumption | How many kWh did I use over the past week and month? |
| Grid import | Grid source | When do I most often pull from the grid? |
| Grid export | PV source / grid | Am I exporting energy at hours when I could be using it myself? |
| PV production | Solar source | What time does production start and end? |
| Devices | Device list on the dashboard | Which device used the most over this period? |
| Water and gas | Separate sources | Do I see an unusual spike in usage? |
Compare the Energy Dashboard against your utility bill, if you have one handy. Differences of a few percent are normal, given different measurement points and different billing periods. A large discrepancy is a signal that something is misconfigured, or that you're measuring at a different point than the utility's billing meter.
Data from your own energy dashboard
The Energy Dashboard shows history. Your own dashboard from Lesson 8 shows today's decisions. In this part of the audit, check whether that dashboard genuinely answers the questions you ask yourself every day.
The "Right now" section: as soon as you open the view, can you tell whether you're importing, exporting, or producing?
The "Today" section: do you have production, import, export, and cost in one place?
Dynamic price and threshold: if you have a pricing integration, do you see the price alongside a reference line, and not just a bare number?
Automations: are the lockout toggles from Lesson 7 visible, and are they actually working?
Your phone: does the dashboard load quickly on a phone, or does it need to be simplified?
If the dashboard isn't helping, say so plainly in the report. "I have too many charts" or "I'm missing a daily cost figure" is a valuable conclusion too. It's better to know what needs fixing than to pretend the cockpit is finished.
Production, import, export, and self-consumption
If you have solar, this part of the audit is the important one. Don't just look at PV production. Look at the balance: how much you used on-site, how much you pulled from the grid, and how much you sent back.
PV production: sensor.pv_production_daily, or the Energy Dashboard's data for the past week.
Import: when the home most often pulls from the grid, especially in the evening and morning.
Export: whether PV surplus goes to the grid at hours when you could have used it locally.
Self-consumption: what percentage of your PV production you used at home rather than exporting.
Self-sufficiency: what percentage of your home's consumption was covered by your own PV.
A practical takeaway
High export paired with low self-consumption means devices are running outside the hours PV is producing. That's not an installation fault. It's a signal that it's worth shifting the water heater, dishwasher, washing machine, or EV charging to daytime, if your comfort allows it.
If you don't have solar, skip this part of the audit. Focus instead on your consumption peaks, and on which devices push your bill up the most.
Your biggest loads
In most homes, a handful of devices account for most of the bill: heating, hot water, the fridge, the oven, the dryer, air conditioning, an EV charger. Home Assistant will only show you this if those devices are measured or visible on the Energy Dashboard.
| Data source | What it gives you | Limitation |
|---|---|---|
| Outlet / Shelly Plug | Accurate measurement of one device | You only see what's plugged into it |
| Energy Dashboard, devices | Ranking of usage over time | Only devices added to the dashboard |
| Whole-home measurement | Import, export, home consumption | Doesn't split out individual devices without extra measurement points |
| Estimate from power rating | A quick guess, e.g. 2000 W for an hour = 2 kWh | It's an approximation, not a billing figure |
In the report, write down the three biggest loads you can actually see in the data. If you only see one device from an outlet, also note two more you suspect based on your bill or how often you use them. That becomes the starting point for a measurement-expansion plan.
Example: a note listing your biggest loads
# Biggest loads: week of Jun 23-29
1. Water heater: 18.4 kWh (sensor.water_heater_energy_daily, week total)
2. Whole-home import: 42 kWh (Energy Dashboard)
3. Office desk: 4.2 kWh (Zigbee outlet)
Conclusion: the water heater mostly runs in the evening, outside PV production hours.Standby and devices running for no good reason
In Lesson 7 we worked out whether standby made financial sense. In this audit, you're hunting for devices that draw power without giving you anything back, or that run at the wrong time of day.
Overnight standby: a desk, a TV, a set-top box, chargers. Check your binary_sensor.*_standby entities from Lesson 4, if you have them.
"Forgotten" devices: a heater in a room nobody uses, a fan, a patio light, an old computer.
Running at peak: a device switched on right when import is highest and PV isn't producing.
False savings: an automation that switches off something drawing 3 W and irritates the household in the process.
Write down one device or scenario that's running without a good reason. It could be standby worth a gentle reminder, or a water heater running overnight during high-import hours. You don't have to fix it right away. First, just name the problem.
Dynamic pricing, your rate plan, and shifting usage
Check data completeness before you draw a conclusion
Check that Home Assistant ran without long gaps during the period you're analyzing, that sensors weren't unavailable, that counters didn't change units, and that all your figures cover the same time range. Missing data can look exactly like savings, when it's really just a gap in measurement.
A week is a test, not always a final verdict
Seven days is enough to check that a sensor and a first automation work. It's not always enough to calculate an annual saving. Weather, temperature, who's home, vacations, guests, an outage, EV charging, and PV production can all shift the result substantially. To properly judge an automation, you also need a baseline period, before the change, under similar conditions and the same rate plan.
Here you're combining three things from Lessons 6 and 7: your rate plan, the dynamic price, and the real windows where you can actually shift something. Home Assistant won't renegotiate your contract for you, but it will show you whether your habits line up with your rate plan and the market.
A flat-rate plan: with a plain flat rate, what you pay per kWh doesn't move with the dynamic price. The dynamic price can still affect the value of what you export, if you're on a net-metering or export-credit arrangement. Shifting usage can still raise your self-consumption even at a fixed purchase price.
A time-of-use plan: check whether the dishwasher, washing machine, or EV charger are actually running in the cheaper windows.
PV plus dynamic pricing: use your own production first, and only then go looking for cheap energy from the market.
Comfort: shifting usage only makes sense if the household can genuinely live with it.
Check your sources
Dynamic pricing rules and net-metering arrangements change over time. This lesson isn't legal or financial advice. Before making a financial decision, check your agreement, your utility, and current, up-to-date sources.
In the report, write down one device you can shift to a cheaper hour, a day with PV surplus, or a low dynamic-price window. If you don't have dynamic pricing or PV, write down one device to shift to the cheaper part of a time-of-use plan instead.
One automation worth keeping
Review the automations from Lessons 4 and 7. Pick one that's working well: it sends a genuinely useful notification, helps you use PV, or reminds you about cheaper energy without annoying the household.
Keep it if: it's run at least a week without false alarms, and has an input_boolean lockout.
Keep it if: you have a rough sense of what it saves, or what it helps you use more efficiently.
Keep it if: the household actually uses it, or at least doesn't switch it off.
Example: you keep the PV-surplus notification for the water heater, because it only fires during the day and never switches the device on without your say-so. Or you keep the cheap-price reminder for the dishwasher, because you start it by hand after the notification arrives.
One automation to disable or fix
Cutting a bad automation matters just as much as keeping a good one. If something's working badly, retire it now. The module's whole point is that fewer sensible automations beat more chaos.
Disable it if: it sends a notification every time a cloud passes over the PV panels.
Disable it if: it controls a device with no measurement and no safeguards.
Fix it if: it has the wrong price threshold, the wrong hours, or no hysteresis.
Fix it if: it saves a handful of cents and ruins comfort in the process.
In the report, write down what you're switching off or fixing, and why. If you don't have any automations yet, write down one thing you're deliberately not automating for now, because you don't have the data yet, or because it's a critical device.
A first 7-day savings plan
A 7-day plan should be small and realistic. Not "I'll cut my bill by 30%." More like: over the course of a week, you do three specific things and check the effect in the data.
| Day | Action | How I'll check the effect |
|---|---|---|
| 1-2 | Keep one good automation / notification running | Note: how many times it helped |
| 3-4 | Shift one device to a cheaper hour or a PV-surplus day | Import / daily cost on the dashboard |
| 5 | Disable or fix one bad automation | Fewer false alarms |
| 6 | Check standby in one spot in the house | A reading from an outlet, or an overnight observation |
| 7 | Compare this week against the last one on the Energy Dashboard | kWh, import, cost, self-consumption |
One week isn't enough for a revolution
It's enough to spot a direction. Weather, guests, vacations, and a different daily rhythm all shift consumption. Don't judge the whole house by one rainy day or one heat wave.
A plan for expanding your measurements
The audit will expose gaps. Maybe you can see the whole home but not the water heater. Maybe you have the water heater but not grid import. A measurement-expansion plan should be an order of priorities, not a shopping list.
Priority 1: whatever has the biggest impact on your bill and your decisions, and still isn't measured.
Priority 2: whole-home measurement, or fixing the import/export direction, if you have PV.
Priority 3: more smart plugs, if they'll pay for themselves within a reasonable time.
Priority 4: a dynamic pricing integration, water, gas, if it would genuinely change your decisions.
Write down one next measurement you want to add, and one device you're not buying right now, because the data shows the savings would be too small. That's a mature decision, not a failure.
A home energy report template
Below is a report template. Copy it into a notebook, Notion, Google Docs, or a plain text file. What matters is filling it in with real numbers from your own home, not generalities.
# Home energy report: Module 17
Date: ____
Period analyzed: last week / last month
## 1. What I have measured
- Energy Dashboard: yes / no / partial
- Own energy dashboard: yes / no
- Device measurement: ____
- PV / import / export: ____
- Dynamic pricing: yes / no
## 2. The key numbers
- Home consumption: ____ kWh
- Import: ____ kWh
- Export: ____ kWh (if applicable)
- PV production: ____ kWh (if applicable)
- Self-consumption: ____ %
- Estimated cost: ____ $
## 3. My three biggest loads
1. ____
2. ____
3. ____
## 4. One problem to fix
- Device / scenario: ____
- Why it's a problem: ____
## 5. Automations
- Keeping: ____
- Disabling / fixing: ____
## 6. One conclusion
The single most important thing I see in the data:
____
## 7. One 7-day plan
1. ____
2. ____
3. ____
## 8. Next measurement to add
____Common mistakes in a closing analysis
A report with no numbers: nothing but generalities like "I use a lot," with no kWh and no cost.
Too many conclusions at once: five plans for the week instead of one sensible one.
Comparing one day to a whole month: a heat wave, guests, or a vacation skews the result.
Ignoring comfort: a savings plan the household simply won't accept.
Buying hardware before the audit: a new smart plug instead of first checking whether the device draws much at all.
Confusing PV with savings by itself: production without self-consumption doesn't always mean a lower bill.
No single conclusion: the report describes everything, but doesn't say what you're doing tomorrow.
Leaving bad automations in place: "maybe it'll work someday," instead of disabling or fixing it.
Assignment: a home energy report
To do
☐ Gather data from the Energy Dashboard for the past week or month.
☐ Open your own dashboard and note what works and what needs simplifying.
☐ If you have PV, record production, import, export, self-consumption, and self-sufficiency.
☐ List the three biggest loads you can see or suspect from the data.
☐ Find one device or scenario that's standby, or otherwise running pointlessly.
☐ Write down one device to shift under your rate plan, PV, or dynamic pricing.
☐ Pick one automation to keep, and one to disable or fix.
☐ Build a 7-day savings plan with three specific actions.
☐ Write down one next measurement to expand the system with.
☐ Fill in the home energy report template.
☐ End the report with one conclusion and one plan of action for tomorrow.
Key takeaways
The mini project ties the whole module together: the Energy Dashboard, your own dashboard, measurements, PV, automations, and costs.
An audit with no numbers is just an opinion: the report needs kWh, percentages, and an approximate cost.
One conclusion beats ten charts: pick the single most important thing the data is telling you.
Keep the good automation, cut the bad one: less chaos, more trust in the system.
A 7-day plan should be small and testable: then you compare one week against the next.
Expanding measurement follows an order: start with whatever changes your decisions the most.
The end of Module 17
You've made the whole journey through the energy module: from asking what's worth measuring, through the Energy Dashboard, entities, costs, your first measurement, whole-home measurement, solar, real savings, and a decision-driven dashboard. In the end, it was never about having the most charts. It's about knowing what your home actually does with energy, and what you're going to do about it.
If you finish this lesson with a report that has one conclusion and a plan for the days ahead, the module has done its job. Home Assistant doesn't save energy on its own. You save it, once the data leads to a decision, and the decision leads to action.
Next up is Module 18: home security in Home Assistant. You'll bring the same habit with you: measure first, understand the data, and only then automate, this time applied to locks, sensors, cameras, and alerts that keep your home safe.
