How to set automatic range for the y axis?

[henriknordgren]

I have a table that show energy consumption in kWh on the y axis. I have set

hours_to_show: 25h
time_offset: 0h

xaxis;
range:
- '2025-03-18 23:30:00'
- '2025-03-19 23:59:00'

Within this interval we have (yet) very little sun, yet the chart look like this:

Why does plotly determine that it should draw the y axis to -8 kWh? We had this production yesterday, but that should not be included, right?

Y axis is set to:

yaxis:
title: Energy (kWh)
tickformat: .1f

[dbuezas]

Post the full yaml (make sure to use code blocks)

[henriknordgren]

You got it maestro!

 type: custom:plotly-graph
            card_mod:
              style: |
                ha-card {         
                  border: none;
                  margin-top: -20px!important;
                }
            hours_to_show: 25h
            time_offset: 0h
            entities:
              - entity: sensor.qubino_3_phase_smart_meter_electric_consumption_kwh
                type: bar
                unit_of_measurement: kWh
                statistic: state
                period: hour
                name: Climate
                marker:
                  color: '#F18303'
                texttemplate: ''
                filters:
                  - delta
              - entity: sensor.easee_lifetime_energy
                type: bar
                unit_of_measurement: kWh
                statistic: state
                period: hour
                name: Easee
                marker:
                  color: '#EA23FB'
                texttemplate: ''
                filters:
                  - delta
              - entity: sensor.ovrigt_lifetime_energy
                type: bar
                unit_of_measurement: kWh
                statistic: state
                period: hour
                name: Övrigt
                marker:
                  color: '#92A0A9'
                texttemplate: ''
                filters:
                  - delta                 
              - entity: sensor.diskmaskin_lifetime_energy
                type: bar
                unit_of_measurement: kWh
                statistic: state
                period: hour
                marker:
                  color: '#92A0A9'
                name: Diskmaskin
                hidden: false
                legendgroup: appliances
                showlegend: false
                texttemplate: ''
                filters:
                  - delta
              - entity: sensor.frysbox_lifetime_energy
                type: bar
                unit_of_measurement: kWh
                statistic: state
                period: hour
                marker:
                  color: '#92A0A9'
                name: Frysbox
                legendgroup: appliances
                showlegend: false
                hidden: true
                texttemplate: ''
                filters:
                  - delta
              - entity: sensor.honshus_lifetime_energy
                type: bar
                unit_of_measurement: kWh
                statistic: state
                hidden: true
                period: hour
                marker:
                  color: '#92A0A9'
                name: Hönshus
                legendgroup: appliances
                texttemplate: ''
                filters:
                  - delta
              - entity: sensor.solar_panels_lifetime_energy
                type: bar
                unit_of_measurement: kWh
                statistic: state
                period: hour
                marker:
                  color: '#FFCC00'
                name: Solar
                texttemplate: ''
                filters:
                  - delta
                  - multiply: -1
            refresh_interval: 30
            config:
              scrollZoom: false
              displayModeBar: false
            layout:
              width: 910
              height: 430
              plot_bgcolor: rgba(0,0,0,0)
              paper_bgcolor: rgba(0,0,0,0)
              font:
                family: Amazon Ember Bold, sans-serif
                size: 14
                color: '#444444'
              barmode: relative
              barcornerradius: 5
              showlegend: false
              legend:
                orientation: h
              xaxis:
                showgrid: false
                dtick: 3600000
                tickformat: '%H'
                tickmode: linear
                tickangle: 0
                range:
                  - '2025-03-18 23:30:00'
                  - '2025-03-19 23:59:00'
              yaxis:
                title: Energy (kWh)
                tickformat: .1f

[dbuezas]

I think it is because of hours_to_show: 25h. Plotly is autoscaling y on the data and then limiting the range of x.

I suggest you use

hours_to_show: current_day
time_offset: -30m

[henriknordgren]

I am still very puzzled about how the hours_to_show and offset actually works. I think this is what I have spent the most time with during the construction of this dashboard.

I have to set it like this:

 hours_to_show: current_day
            time_offset: -65m

this actually looks good visually, but the offset can't be 30m but > 60m. Otherwise I get no bar for hours 00-01, which has been bugging me for a long time.

The above works fine for today, I can with that code see 00-01

But for yesterday, if I set like this:

   hours_to_show: current_day
            time_offset: -1510m

where time_offset: -(60*24)-70) minus one day > translate to minutes, and then the midnight offset of -70 minutes

As you now can see, I am missing the hour 23-00 of yesterday.

EIther I am totally misunderstanding how this works, or something strange is going on.

[dbuezas]

I see. The delta filter is the issue here: it cannot generate an output for the first datapoint because that one doesn't have a previous one to do the subtraction.
Small tip: you can do this for clarity time_offset: $ex (-60*24)-70) + 'm'

hours_to_show: curent_xxx calculates how much to fetch from the past so that the current_xxx (day, hour, month...) is covered. it then also extends the x range so you can see the remaining time to cover the full xxx (day, hour, ...)
time_offset: xxx just shifts the visible part around before fetching. I suggest you revisit their explanations in the readme, there's more info there

• https://github.com/dbuezas/lovelace-plotly-graph-card?tab=readme-ov-file#offsets
• https://github.com/dbuezas/lovelace-plotly-graph-card?tab=readme-ov-file#hours_to_show

[henriknordgren]

Ok, it slowly starts to make sense. If no data exist for hour 23 the day before, then the delta for 00-01 can't be calculated. So a "better" way to do this would be to get data within a certain interval to make sure we get 25 hours for each day, where the first hour is 23-00 the day before? I guess that is why I set hours_to_show: 25h in the first place.

Time offset I set to current hour -(0-24) - (no of days back in time * 24)

So, if the time right now is 22:21 and I want to look at today´s data, my calculation would equal to an offset of -22h and yesterday´s data would equal to an offset of -22 - (1*24) = -46h

Does this makes any kind of sense?

WIth this example, I still dont get why the y axis scales to -8 kWh, if the "window" I am looking at doesn't have these values?

[dbuezas]

Right, delta is calculated as result[i] = data[i]-data[i-1]. So you need an extra datapoint to the left.

For the autoscale, try not settting layout/xaxis/range. Just do

hours_to_show: current_day
time_offset: $ex `${-5 - 0.01}d` # -5 is the days in the past, the -0.01 just adds a tiny bit of the past so the delta has something to grab to for the first datapoint