Plot year by year in the same plot (plotly)

I tried plotting time series with many years in the same start and end day-month.

For example, what I need is data from 01/01 for 2018, 2019, etc in the same plot in order to compare different data from different years.

The code that I do, plot year by year in subplot, but I would like a single plotly plot.

Link of data to download and can run the script

from datetime import date, datetime

import numpy as np
import pandas as pd
import plotly.express as px
import plotly.graph_objects as go

df_decremento_municipio = pd.read_csv('app/data/decremento_municipio_202305291512.csv', index_col="view_date") 
template_graph = {
    "layout": {
        "modebar": {
            "remove": [
                "zoom",
                "pan",
                "select",
                "zoomIn",
                "zoomOut",
                "lasso2d",
                "autoscale",
            ]
        },
        "separators": ".",
        "showlegend": True,
    }
}

df_decremento_municipio.index = pd.to_datetime(df_decremento_municipio.index)
df_decremento_municipio["year"] = df_decremento_municipio.index.year

min_date = df_decremento_municipio.index.date.min()
max_date = df_decremento_municipio.index.date.max()

df=pd.Series(name="area_ha", dtype="float64")
for ano in df_decremento_municipio.index.year.unique()[
    df_decremento_municipio.index.year.unique() > 2016
]:
    # Remover os anos de 2015 e 2016 (dados muito ruins)

    df_por_ano = df_decremento_municipio[df_decremento_municipio["year"] == ano]
    dff_acumulacao = df_por_ano["area_ha"].groupby([df_por_ano.index]).sum().cumsum()
    df = df.append(dff_acumulacao)

fig = go.Figure()

# Primeira data de cada dado no ano
start = ["2020-01-01", "2018-01-04", "2021-01-05" , "2022-01-05", "2019-01-01", "2023-01-10", "2017-12-27"]
# Ultima data de cada dado no ano
end =   ["2020-12-31", "2018-12-27", "2021-12-16", "2022-12-31", "2019-12-22",  "2023-05-03", "2017-01-26"]

years = df.index.year.unique()[df.index.year.unique()>2016].sort_values()

for idx, (s,e) in enumerate(zip(start, end)):
    tmp = df[(df.index >= start[idx]) & (df.index <= end[idx])]
    fig.add_trace(go.Scatter(x=tmp.index,
                             y=tmp,
                             name=str(years[idx]),
                             mode='lines',
                            ))

fig.update_layout(height=600, xaxis_tickformat='%d-%m')
fig.update_xaxes(type='date')

fig.show()

Since your x-axis contains the information about the year, the lines in your plot will be discontinuous (as each line will start where the previous one left off). To hide the year information using xaxis_tickformat='%d-%m' will not help in this case, it will only change the visuals.

What you can do to compare the time series across years is to simply remove the information about the year from the plot. In your add_trace, change the x-axis values to be only the month and day using strftime('%m-%d'):

fig.add_trace(go.Scatter(x=tmp.index.strftime('%m-%d'),
                             y=tmp['dist'],
                             name=str(years[idx]),
                             mode='lines',
                            ))

Note that you will also need to remove the fig.update_xaxes(type='date') line as the data type is now strings.

What I've done, and I belive that is more rational than the suggestion in answer gone in https://stackoverflow.com/questions/69013744/a-simple-way-to-plot-day-and-month-only-on-the-x-axis-to-compare-years

I've created a time lapse column and put it in x axes.

Plotly doesn't work well with column with type of column as "timelapse". But, I think it is better that I can do.

import pandas as pd
import plotly.express as px

df_decremento_municipio = pd.read_csv('app/graficos_dev/data/decremento_municipio_202305291512.csv', index_col="view_date") 

template_graph = {
    "layout": {
        "modebar": {
            "remove": [
                "zoom",
                "pan",
                "select",
                "zoomIn",
                "zoomOut",
                "lasso2d",
                "autoscale",
            ]
        },
        "separators": ".",
        "showlegend": True,
    }
}

df_decremento_municipio.index = pd.to_datetime(df_decremento_municipio.index)

df_decremento_municipio = df_decremento_municipio[["area_ha"]].groupby(df_decremento_municipio.index).sum()

t1 = pd.DataFrame()

for ano in df_decremento_municipio.index.year.unique()[
    df_decremento_municipio.index.year.unique() > 2017
]:  

    df = df_decremento_municipio[df_decremento_municipio.index.year == ano]
    df["year"] = df.index.year
    df["timedelta"] = (df.index - (df.index.year.astype("str") + "-01-01").astype("datetime64[ns]"))/1000000
    df["cumsum"]= df["area_ha"].cumsum()

    t1 = pd.concat([t1, df[["year", "timedelta", "cumsum"]]])

fig = px.line(t1, x="timedelta", y="cumsum", color='year')

fig.update_layout(title="Desflorestamento por Tempo",
    xaxis={"title": "Data"},
    yaxis={"title": "Área (ha)"},
    xaxis_tickformat = '%d-%m'
)

fig.update_xaxes(type='date')

fig.show()

The result of figure: