Stacking bar chart with total_increasing sensor

[ando2040]

I'm so close on this one, any ideas please?

After a few bad ideas ChatGPT did come up with barmode:stack instead of stackgroup, and also suggested derivate without the d which I found myself in the readme. Two of the series are stacking correctly, but the climate/heat pump one is not. So close on the 21st (heat pump value is within 0.7), but incorrect negative values on most of the other days - 15th shown should be positive 8.6.

These are all state_class: total_increasing sensors so I thought statistic: state would be appropriate reading the notes. However, sum works properly for two, and makes no difference for climate!

Climate happens to be a custom sensor so I could change it... but history-explorer-card extracts the daily values correctly and consistently for all three sensors.

type: custom:plotly-graph
layout:
  barmode: stack
entities:
  - entity: sensor.immersion_water_heater_daily_consumption
    name: Immersion
    type: bar
    statistic: sum
    period: day
    filters:
      - derivate: d
  - entity: sensor.climate_total_daily_electrical_consumption
    name: Heat Pump
    type: bar
    statistic: state
    period: day
    filters:
      - derivate: d
  - entity: sensor.myenergi_zappi_xxxxxxxx_energy_used_today_2
    name: EV
    type: bar
    statistic: sum
    period: day
    filters:
      - derivate: d
hours_to_show: 300
refresh_interval: 10

[dbuezas]

What is it that is not working as you want?

[dbuezas]

The derivative is the rate of change, per day in your case (as you configured it). Since the source sensor was less than the day before, the derivative is negative.
What exactly are you going after? maybe I can point to a solution

[ando2040]

I'm looking for the positive change since the previous day (8.6kWh on 15th for example - it's a total_increasing sensor, I never expect a negative value.)

But I tried a delta filter and that was unsurprisingly rendered on a separate y axis, and the values were still (mostly) negative. I also don't really understand what's different about this sensor vs the other two.

[dbuezas]

then don't use derivate. Just plot the data as is

[ando2040]

David, thanks for making me go back to the beginning.

For the record, below what I was aiming for. I started down a filters route because I decided delta between two days of the total_increasing sensors was what I needed, then the bot sent me to derivate and I didn't engage my brain. It turns out statistic: state was right. I think I only dismissed it because I came to it late and didn't remove either derivate or delta filter when I tried it... also because derivate was producing an answer deceptively close to desired for two of three sensors. Bots are very bad at looking holistically at a problem and changing more than one thing, I've experienced that before.

type: custom:plotly-graph
layout:
  barmode: stack
entities:
  - entity: sensor.immersion_water_heater_daily_consumption
    name: Immersion
    type: bar
    statistic: state
    period: day
    filters:
      - filter: i > 0
  - entity: sensor.climate_total_daily_electrical_consumption
    name: Heat Pump
    type: bar
    statistic: state
    period: day
    filters:
      - filter: i > 0
  - entity: sensor.myenergi_zappi_xxxxxxxx_energy_used_today_2
    name: EV
    type: bar
    statistic: state
    period: day
    filters:
      - filter: i > 0
hours_to_show: 360
refresh_interval: 10