Skip to content

edsteva.viz.plots

probe_plot 

probe_plot(
    probe: BaseProbe,
    fitted_model: BaseModel = None,
    care_site_level: List[str] = None,
    start_date: datetime = None,
    end_date: datetime = None,
    save_path: str = None,
    legend_predictor: str = "Predictor c(t)",
    legend_model: str = "Model f(t)",
    x_axis_title: str = None,
    y_axis_title: str = None,
    main_chart_config: Dict[str, float] = None,
    model_line_config: Dict[str, str] = None,
    probe_line_config: Dict[str, str] = None,
    chart_style: Dict[str, float] = None,
    indexes_to_remove: List[str] = None,
    **kwargs
)

Displays a chart with the average completeness predictor \(c(t)\) over time \(t\) with the fitted model \(\hat{c}(t)\) if specified. The chart is exportable in png or svg format and easy to integrate into a report. Is also possible to save the chart in HTML with the "save_path" optional input.

PARAMETER DESCRIPTION
probe

Class describing the completeness predictor \(c(t)\).

TYPE: BaseProbe

fitted_model

Model fitted to the probe

TYPE: BaseModel DEFAULT: None

care_site_level

EXAMPLE: ["Hospital"], ["Hôpital", "UF"] or ["UF", "UH"]

TYPE: List[str] DEFAULT: None

start_date

EXAMPLE: "2019-05-01"

TYPE: datetime DEFAULT: None

end_date

EXAMPLE: "2021-07-01"

TYPE: datetime DEFAULT: None

save_path

Folder path where to save the chart in HTML format.

EXAMPLE: "my_folder/my_file.html"

TYPE: str DEFAULT: None

legend_predictor

Label name for the predictor legend.

TYPE: str DEFAULT: 'Predictor c(t)'

legend_model

Label name for the model legend.

TYPE: str DEFAULT: 'Model f(t)'

x_axis_title

Label name for the x axis.

TYPE: str DEFAULT: None

y_axis_title

Label name for the y axis.

TYPE: str DEFAULT: None

main_chart_config

If not None, configuration used to construct the top main chart.

TYPE: Dict[str, float] DEFAULT: None

model_line_config

If not None, configuration used to construct the model line.

TYPE: Dict[str, str] DEFAULT: None

probe_line_config

If not None, configuration used to construct the probe line.

TYPE: Dict[str, str] DEFAULT: None

chart_style

If not None, configuration used to configure the chart style.

EXAMPLE: {"labelFontSize": 13, "titleFontSize": 14}

TYPE: Dict[str, float] DEFAULT: None

indexes_to_remove

indexes to remove from the groupby selection.

TYPE: List[str] DEFAULT: None

Source code in edsteva/viz/plots/probe/wrapper.py 
 19
 20
 21
 22
 23
 24
 25
 26
 27
 28
 29
 30
 31
 32
 33
 34
 35
 36
 37
 38
 39
 40
 41
 42
 43
 44
 45
 46
 47
 48
 49
 50
 51
 52
 53
 54
 55
 56
 57
 58
 59
 60
 61
 62
 63
 64
 65
 66
 67
 68
 69
 70
 71
 72
 73
 74
 75
 76
 77
 78
 79
 80
 81
 82
 83
 84
 85
 86
 87
 88
 89
 90
 91
 92
 93
 94
 95
 96
 97
 98
 99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
133
134
135
136
137
138
139
140
141
142
143
144
145
146
147
def probe_plot(
    probe: BaseProbe,
    fitted_model: BaseModel = None,
    care_site_level: List[str] = None,
    start_date: datetime = None,
    end_date: datetime = None,
    save_path: str = None,
    legend_predictor: str = "Predictor c(t)",
    legend_model: str = "Model f(t)",
    x_axis_title: str = None,
    y_axis_title: str = None,
    main_chart_config: Dict[str, float] = None,
    model_line_config: Dict[str, str] = None,
    probe_line_config: Dict[str, str] = None,
    chart_style: Dict[str, float] = None,
    indexes_to_remove: List[str] = None,
    **kwargs,
):
    r"""
    Displays a chart with the average completeness predictor $c(t)$ over time $t$ with the fitted model $\hat{c}(t)$ if specified.
    The chart is exportable in png or svg format and easy to integrate into a report. Is also possible to save the chart in HTML with the "save_path" optional input.

    Parameters
    ----------
    probe : BaseProbe
        Class describing the completeness predictor $c(t)$.
    fitted_model : BaseModel, optional
        Model fitted to the probe
    care_site_level : List[str], optional
        **EXAMPLE**: `["Hospital"]`, `["Hôpital", "UF"]` or `["UF", "UH"]`
    start_date : datetime, optional
        **EXAMPLE**: `"2019-05-01"`
    end_date : datetime, optional
        **EXAMPLE**: `"2021-07-01"`
    save_path : str, optional
        Folder path where to save the chart in HTML format.

        **EXAMPLE**: `"my_folder/my_file.html"`
    legend_predictor: str, optional,
        Label name for the predictor legend.
    legend_model: str, optional,
        Label name for the model legend.
    x_axis_title: str, optional,
        Label name for the x axis.
    y_axis_title: str, optional,
        Label name for the y axis.
    main_chart_config: Dict[str, str], optional
        If not None, configuration used to construct the top main chart.
    model_line_config: Dict[str, str], optional
        If not None, configuration used to construct the model line.
    probe_line_config: Dict[str, str], optional
        If not None, configuration used to construct the probe line.
    chart_style: Dict[str, float], optional
        If not None, configuration used to configure the chart style.

        **EXAMPLE**: `{"labelFontSize": 13, "titleFontSize": 14}`
    indexes_to_remove: List[str], optional
        indexes to remove from the groupby selection.
    """
    alt.data_transformers.enable("default")
    alt.data_transformers.disable_max_rows()

    probe_config = deepcopy(probe.get_viz_config("probe_plot"))
    if main_chart_config is None:
        main_chart_config = probe_config["main_chart"]
    if chart_style is None:
        chart_style = probe_config["chart_style"]
    indexes = get_indexes_to_groupby(
        predictor_columns=probe.predictor.columns,
        predictor_metrics=probe._metrics.copy(),
        indexes_to_remove=indexes_to_remove,
    )

    if fitted_model:
        predictor = fitted_model.predict(probe).copy()
    else:
        predictor = probe.predictor.copy()

    predictor = filter_data(
        data=predictor,
        care_site_level=care_site_level,
        start_date=start_date,
        end_date=end_date,
        **kwargs,
    )

    indexes = [
        {
            "field": variable["field"],
            "title": variable["field"].replace("_", " ").capitalize(),
        }
        for variable in indexes
        if variable["field"] in predictor.columns
        and len(predictor[variable["field"]].unique()) >= 2
    ]

    if fitted_model:
        model_config = deepcopy(fitted_model.get_viz_config("probe_plot"))
        if model_line_config is None:
            model_line_config = model_config["model_line"]
        if probe_line_config is None:
            probe_line_config = model_config["probe_line"]
        chart = fitted_probe_line(
            predictor=predictor,
            indexes=indexes,
            legend_predictor=legend_predictor,
            legend_model=legend_model,
            x_axis_title=x_axis_title,
            y_axis_title=y_axis_title,
            main_chart_config=main_chart_config,
            model_line_config=model_line_config,
            probe_line_config=probe_line_config,
        )
    else:
        chart = probe_line(
            predictor=predictor,
            indexes=indexes,
            x_axis_title=x_axis_title,
            y_axis_title=y_axis_title,
            main_chart_config=main_chart_config,
        )

    if save_path:
        save_html(
            obj=configure_style(chart=chart, chart_style=chart_style),
            filename=save_path,
        )

    return chart

estimates_densities_plot 

estimates_densities_plot(
    fitted_model: BaseModel,
    probe: BaseProbe = None,
    care_site_level: List[str] = None,
    start_date: Union[datetime, str] = None,
    end_date: Union[datetime, str] = None,
    save_path: str = None,
    vertical_bar_charts_config: Dict[str, str] = None,
    horizontal_bar_charts_config: Dict[str, str] = None,
    chart_style: Dict[str, float] = None,
    y_axis_title: str = None,
    **kwargs
)

Displays the density plot with the associated box plot of each estimate and metric computed in the input model. It can help you to set the thresholds.

PARAMETER DESCRIPTION
probe

Class describing the completeness predictor \(c(t)\)

TYPE: BaseProbe DEFAULT: None

fitted_model

Model with estimates of interest.

EXAMPLE: StepFunction Model with \((\hat{t_0}, \hat{c_0})\)

TYPE: BaseModel

care_site_level

EXAMPLE: ["Hospital"], ["Hôpital", "UF"] or ["UF", "UH"]

TYPE: List[str] DEFAULT: None

start_date

EXAMPLE: "2019-05-01"

TYPE: datetime DEFAULT: None

end_date

EXAMPLE: "2021-07-01"

TYPE: datetime DEFAULT: None

save_path

Folder path where to save the chart in HTML format.

TYPE: str DEFAULT: None

vertical_bar_charts_config

Configuration used to construct the vertical bar charts.

TYPE: Dict[str, str] DEFAULT: None

horizontal_bar_charts_config

Configuration used to construct the horizontal bar charts.

TYPE: Dict[str, str] DEFAULT: None

chart_style

Configuration used to configure the chart style.

EXAMPLE: {"labelFontSize": 13, "titleFontSize": 14}

TYPE: Dict[str, float] DEFAULT: None

y_axis_title

Label name for the y axis.

TYPE: str DEFAULT: None

Source code in edsteva/viz/plots/estimates_densities/estimates_densities.py 
 21
 22
 23
 24
 25
 26
 27
 28
 29
 30
 31
 32
 33
 34
 35
 36
 37
 38
 39
 40
 41
 42
 43
 44
 45
 46
 47
 48
 49
 50
 51
 52
 53
 54
 55
 56
 57
 58
 59
 60
 61
 62
 63
 64
 65
 66
 67
 68
 69
 70
 71
 72
 73
 74
 75
 76
 77
 78
 79
 80
 81
 82
 83
 84
 85
 86
 87
 88
 89
 90
 91
 92
 93
 94
 95
 96
 97
 98
 99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
133
134
135
136
137
138
139
140
141
142
143
144
145
146
147
148
149
150
151
152
153
154
155
156
157
158
159
160
161
162
163
164
165
166
167
168
169
170
171
172
173
174
175
176
177
178
179
180
181
182
183
184
185
186
187
188
189
190
191
192
193
194
195
196
197
198
199
200
201
202
203
204
205
206
207
208
209
210
211
212
213
214
215
216
217
218
219
220
221
222
223
224
225
226
227
228
229
230
def estimates_densities_plot(
    fitted_model: BaseModel,
    probe: BaseProbe = None,
    care_site_level: List[str] = None,
    start_date: Union[datetime, str] = None,
    end_date: Union[datetime, str] = None,
    save_path: str = None,
    vertical_bar_charts_config: Dict[str, str] = None,
    horizontal_bar_charts_config: Dict[str, str] = None,
    chart_style: Dict[str, float] = None,
    y_axis_title: str = None,
    **kwargs,
):
    r"""Displays the density plot with the associated box plot of each estimate and metric computed in the input model. It can help you to set the thresholds.


    Parameters
    ----------
    probe : BaseProbe
        Class describing the completeness predictor $c(t)$
    fitted_model : BaseModel
        Model with estimates of interest.

        **EXAMPLE**: StepFunction Model with $(\hat{t_0}, \hat{c_0})$
    care_site_level : List[str], optional
        **EXAMPLE**: `["Hospital"]`, `["Hôpital", "UF"]` or `["UF", "UH"]`
    start_date : datetime, optional
        **EXAMPLE**: `"2019-05-01"`
    end_date : datetime, optional
        **EXAMPLE**: `"2021-07-01"`
    save_path : str, optional
        Folder path where to save the chart in HTML format.
    vertical_bar_charts_config: Dict[str, str], optional
        Configuration used to construct the vertical bar charts.
    horizontal_bar_charts_config: Dict[str, str], optional
        Configuration used to construct the horizontal bar charts.
    chart_style: Dict[str, float], optional
        Configuration used to configure the chart style.

        **EXAMPLE**: `{"labelFontSize": 13, "titleFontSize": 14}`
    y_axis_title: str, optional,
        Label name for the y axis.
    """
    alt.data_transformers.disable_max_rows()

    estimates = fitted_model.estimates.copy()
    estimates = filter_data(
        data=estimates,
        table_name="estimates",
        care_site_level=care_site_level,
        **kwargs,
    )
    if probe is not None:
        predictor = probe.predictor.copy()
        # Filter data in predictor not in estimates
        predictor = predictor.merge(
            estimates[list(estimates.columns.intersection(set(predictor.columns)))],
            on=list(estimates.columns.intersection(set(predictor.columns))),
        )
        predictor = filter_data(
            data=predictor,
            start_date=start_date,
            end_date=end_date,
        )
        estimates = probe.add_names_columns(estimates)
        probe_config = deepcopy(probe.get_viz_config("estimates_densities_plot"))
        if vertical_bar_charts_config is None:
            vertical_bar_charts_config = probe_config["vertical_bar_charts"]
        if horizontal_bar_charts_config is None:
            horizontal_bar_charts_config = probe_config["horizontal_bar_charts"]
        if chart_style is None:
            chart_style = probe_config["chart_style"]

    quantitative_estimates = []
    time_estimates = []

    base_estimate = alt.Chart(estimates)
    for estimate in fitted_model._coefs + fitted_model._metrics:
        if estimates[estimate].dtype == float or estimates[estimate].dtype == int:
            max_value = estimates[estimate].max()
            min_value = estimates[estimate].min()
            estimates[estimate] = round(estimates[estimate], 3)

            estimate_density = (
                alt.vconcat(
                    (
                        (
                            base_estimate.transform_density(
                                estimate,
                                as_=[estimate, "Density"],
                                extent=[min_value, max_value],
                            )
                            .mark_area()
                            .encode(
                                x=alt.X("{}:Q".format(estimate), title=None),
                                y=alt.Y("Density:Q", title=y_axis_title),
                            )
                        )
                        + base_estimate.mark_rule(color="red").encode(
                            x="median({}):Q".format(estimate),
                            tooltip=alt.Tooltip("median({}):Q".format(estimate)),
                        )
                    ).properties(width=800, height=300),
                    (
                        base_estimate.mark_tick().encode(
                            x=alt.X("{}:Q".format(estimate), axis=None)
                        )
                    ),
                    spacing=0,
                )
                & (
                    base_estimate.mark_boxplot().encode(
                        x="{}:Q".format(estimate),
                    )
                )
            ).resolve_scale(x="shared")
            quantitative_estimates.append(estimate_density)

        else:
            estimates[estimate] = estimates[estimate].astype("datetime64[ns]")
            estimate_density = (
                (
                    base_estimate.transform_timeunit(
                        estimate="yearmonth({})".format(estimate)
                    )
                    .mark_bar(size=10)
                    .encode(
                        x=alt.X(
                            "{}:T".format(estimate),
                            axis=alt.Axis(
                                tickCount="month",
                                format="%Y, %b",
                                labelAngle=-90,
                            ),
                            title=estimate,
                        ),
                        y=alt.Y(
                            "count({}):Q".format(estimate),
                            axis=alt.Axis(tickMinStep=1),
                            title=y_axis_title,
                        ),
                    )
                )
                + base_estimate.mark_rule(color="red").encode(
                    x="median({}):T".format(estimate),
                    tooltip=alt.Tooltip("median({}):T".format(estimate)),
                )
            ).properties(width=800, height=300)
            time_estimates.append(estimate_density)

    estimates_densities = time_estimates + quantitative_estimates

    chart = reduce(
        lambda estimate_density_1, estimate_density_2: estimate_density_1
        & estimate_density_2,
        estimates_densities,
    )
    if probe is not None:
        base = alt.Chart(predictor)

        horizontal_bar_charts, y_variables_selections = generate_horizontal_bar_charts(
            base=base,
            horizontal_bar_charts_config=horizontal_bar_charts_config,
            predictor=predictor,
        )
        vertical_bar_charts, x_variables_selections = generate_vertical_bar_charts(
            base=base,
            vertical_bar_charts_config=vertical_bar_charts_config,
            predictor=predictor,
        )

        selections = dict(
            y_variables_selections,
            **x_variables_selections,
        )
        selection_charts = dict(
            horizontal_bar_charts,
            **vertical_bar_charts,
        )
        chart = add_interactive_selection(
            base=chart, selection_charts=selection_charts, selections=selections
        )

        chart = concatenate_charts(
            main_chart=chart,
            horizontal_bar_charts=horizontal_bar_charts,
            vertical_bar_charts=vertical_bar_charts,
            spacing=0,
        )
    elif "care_site_level" in estimates.columns:
        care_site_level_dropdwon = alt.binding_select(
            options=estimates["care_site_level"].unique(), name="Care site level : "
        )
        care_site_level_selection = alt.selection_point(
            fields=["care_site_level"],
            bind=care_site_level_dropdwon,
            value=estimates["care_site_level"].unique()[0],
        )
        chart = chart.add_params(care_site_level_selection).transform_filter(
            care_site_level_selection
        )
    if chart_style:
        chart = configure_style(chart=chart, chart_style=chart_style)
    if save_path:
        save_html(
            obj=chart,
            filename=save_path,
        )

    return chart

normalized_probe_plot 

normalized_probe_plot(
    probe: BaseProbe,
    fitted_model: BaseModel,
    care_site_level: List[str] = None,
    start_date: Union[datetime, str] = None,
    end_date: Union[datetime, str] = None,
    t_min: int = None,
    t_max: int = None,
    save_path: str = None,
    x_axis_title: str = None,
    y_axis_title: str = None,
    main_chart_config: Dict[str, float] = None,
    model_line_config: Dict[str, str] = None,
    probe_line_config: Dict[str, str] = None,
    error_line_config: Dict[str, str] = None,
    estimates_selections: Dict[str, str] = None,
    estimates_filters: Dict[str, str] = None,
    chart_style: Dict[str, float] = None,
    indexes_to_remove: List[str] = ["care_site_id"],
    **kwargs
)

Displays a chart with the aggregated normalized completeness predictor \(\frac{c(\Delta t)}{c_0}\) over normalized time \(\Delta t = t - t_0\). It represents the overall deviation from the Model.

Is is possible to save the chart in HTML with the "save_path" optional input.

PARAMETER DESCRIPTION
probe

Class describing the completeness predictor \(c(t)\)

TYPE: BaseProbe

fitted_model

Model fitted to the probe

TYPE: BaseModel

care_site_level

EXAMPLE: ["Hospital"], ["Hôpital", "UF"] or ["UF", "UH"]

TYPE: List[str] DEFAULT: None

start_date

EXAMPLE: "2019-05-01"

TYPE: datetime DEFAULT: None

end_date

EXAMPLE: "2021-07-01"

TYPE: datetime DEFAULT: None

t_min

Minimal difference with \(t_0\) in month \(\Delta t_{min}\).

EXAMPLE: -24

TYPE: int DEFAULT: None

t_max

Maximal difference with \(t_0\) in month \(\Delta t_{max}\).

EXAMPLE: 24

TYPE: int DEFAULT: None

save_path

Folder path where to save the chart in HTML format.

TYPE: str DEFAULT: None

x_axis_title

Label name for the x axis.

TYPE: str DEFAULT: None

y_axis_title

Label name for the y axis.

TYPE: str DEFAULT: None

main_chart_config

If not None, configuration used to construct the top main chart.

TYPE: Dict[str, float] DEFAULT: None

model_line_config

If not None, configuration used to construct the model line.

TYPE: Dict[str, str] DEFAULT: None

error_line_config

If not None, configuration used to construct the error line.

TYPE: Dict[str, str] DEFAULT: None

probe_line_config

If not None, configuration used to construct the probe line.

TYPE: Dict[str, str] DEFAULT: None

estimates_selections

If not None, configuration used to construct the estimates selections.

TYPE: Dict[str, str] DEFAULT: None

estimates_filters

If not None, configuration used to construct the estimates filters.

TYPE: Dict[str, str] DEFAULT: None

chart_style

If not None, configuration used to configure the chart style.

EXAMPLE: {"labelFontSize": 13, "titleFontSize": 14}

TYPE: Dict[str, float] DEFAULT: None

indexes_to_remove

indexes to remove from the groupby selection.

TYPE: List[str] DEFAULT: ['care_site_id']

Source code in edsteva/viz/plots/normalized_probe/normalized_probe.py 
 25
 26
 27
 28
 29
 30
 31
 32
 33
 34
 35
 36
 37
 38
 39
 40
 41
 42
 43
 44
 45
 46
 47
 48
 49
 50
 51
 52
 53
 54
 55
 56
 57
 58
 59
 60
 61
 62
 63
 64
 65
 66
 67
 68
 69
 70
 71
 72
 73
 74
 75
 76
 77
 78
 79
 80
 81
 82
 83
 84
 85
 86
 87
 88
 89
 90
 91
 92
 93
 94
 95
 96
 97
 98
 99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
133
134
135
136
137
138
139
140
141
142
143
144
145
146
147
148
149
150
151
152
153
154
155
156
157
158
159
160
161
162
163
164
165
166
167
168
169
170
171
172
173
174
175
176
177
178
179
180
181
182
183
184
185
186
187
188
189
190
191
192
193
194
195
196
197
198
199
200
201
def normalized_probe_plot(
    probe: BaseProbe,
    fitted_model: BaseModel,
    care_site_level: List[str] = None,
    start_date: Union[datetime, str] = None,
    end_date: Union[datetime, str] = None,
    t_min: int = None,
    t_max: int = None,
    save_path: str = None,
    x_axis_title: str = None,
    y_axis_title: str = None,
    main_chart_config: Dict[str, float] = None,
    model_line_config: Dict[str, str] = None,
    probe_line_config: Dict[str, str] = None,
    error_line_config: Dict[str, str] = None,
    estimates_selections: Dict[str, str] = None,
    estimates_filters: Dict[str, str] = None,
    chart_style: Dict[str, float] = None,
    indexes_to_remove: List[str] = ["care_site_id"],
    **kwargs,
):
    r"""Displays a chart with the aggregated normalized completeness predictor $\frac{c(\Delta t)}{c_0}$ over normalized time $\Delta t = t - t_0$. It represents the overall deviation from the Model.

    Is is possible to save the chart in HTML with the "save_path" optional input.

    Parameters
    ----------
    probe : BaseProbe
        Class describing the completeness predictor $c(t)$
    fitted_model : BaseModel
        Model fitted to the probe
    care_site_level : List[str], optional
        **EXAMPLE**: `["Hospital"]`, `["Hôpital", "UF"]` or `["UF", "UH"]`
    start_date : datetime, optional
        **EXAMPLE**: `"2019-05-01"`
    end_date : datetime, optional
        **EXAMPLE**: `"2021-07-01"`
    t_min : int, optional
        Minimal difference with $t_0$ in month $\Delta t_{min}$.

        **EXAMPLE**: `-24`
    t_max : int, optional
        Maximal difference with $t_0$ in month $\Delta t_{max}$.

        **EXAMPLE**: `24`
    save_path : str, optional
        Folder path where to save the chart in HTML format.
    x_axis_title: str, optional,
        Label name for the x axis.
    y_axis_title: str, optional,
        Label name for the y axis.
    main_chart_config: Dict[str, str], optional
        If not None, configuration used to construct the top main chart.
    model_line_config: Dict[str, str], optional
        If not None, configuration used to construct the model line.
    error_line_config: Dict[str, str], optional
        If not None, configuration used to construct the error line.
    probe_line_config: Dict[str, str], optional
        If not None, configuration used to construct the probe line.
    estimates_selections: Dict[str, str], optional
        If not None, configuration used to construct the estimates selections.
    estimates_filters: Dict[str, str], optional
        If not None, configuration used to construct the estimates filters.
    chart_style: Dict[str, float], optional
        If not None, configuration used to configure the chart style.

        **EXAMPLE**: `{"labelFontSize": 13, "titleFontSize": 14}`
    indexes_to_remove: List[str], optional
        indexes to remove from the groupby selection.
    """

    alt.data_transformers.disable_max_rows()

    # Pre-processing
    predictor = probe.predictor.copy()
    predictor_metrics = probe._metrics.copy()
    estimates = fitted_model.estimates.copy()
    indexes = get_indexes_to_groupby(
        predictor_columns=predictor.columns,
        predictor_metrics=predictor_metrics,
        indexes_to_remove=indexes_to_remove,
    )
    predictor = predictor.merge(estimates, on=probe._index)
    predictor["normalized_date"] = month_diff(
        predictor["date"], predictor["t_0"]
    ).astype(int)
    for estimate in fitted_model._coefs + fitted_model._metrics:
        if pd.api.types.is_datetime64_any_dtype(predictor[estimate]):
            predictor[estimate] = predictor[estimate].dt.strftime("%Y-%m")
    predictor["normalized_c"] = predictor["c"].where(
        (predictor["normalized_date"] < 0) | (predictor["c_0"] == 0),
        predictor["c"] / predictor["c_0"],
    )
    predictor["model"] = 1
    predictor["model"] = predictor["model"].where(predictor["normalized_date"] >= 0, 0)
    predictor = filter_data(
        data=predictor,
        care_site_level=care_site_level,
        start_date=start_date,
        end_date=end_date,
        **kwargs,
    )
    if t_min:
        predictor = predictor[predictor.normalized_date >= t_min]
    if t_max:
        predictor = predictor[predictor.normalized_date <= t_max]

    # Get viz config
    probe_config = deepcopy(probe.get_viz_config("normalized_probe_plot"))
    model_config = deepcopy(
        fitted_model.get_viz_config("normalized_probe_plot", predictor=predictor)
    )
    if probe_line_config is None:
        probe_line_config = model_config["probe_line"]
    if model_line_config is None:
        model_line_config = model_config["model_line"]
    if error_line_config is None:
        error_line_config = model_config["error_line"]
    if estimates_selections is None:
        estimates_selections = model_config["estimates_selections"]
    if estimates_filters is None:
        estimates_filters = model_config["estimates_filters"]
    if main_chart_config is None:
        main_chart_config = probe_config["main_chart"]
    if chart_style is None:
        chart_style = probe_config["chart_style"]

    # Viz
    predictor["legend_model"] = (
        model_line_config.get("legend_title")
        if model_line_config.get("legend_title")
        else type(fitted_model).__name__
    )
    predictor["legend_predictor"] = probe_line_config["legend_title"]
    predictor["legend_error_band"] = error_line_config["legend_title"]
    index_selection, index_fields = create_groupby_selection(
        indexes=indexes,
        predictor=predictor,
    )
    base = alt.Chart(predictor)
    base = add_estimates_filters(
        base=base,
        estimates_filters=estimates_filters,
    )
    main_chart = generate_main_chart(
        base=base,
        main_chart_config=main_chart_config,
        index_selection=index_selection,
        index_fields=index_fields,
        x_axis_title=x_axis_title,
        y_axis_title=y_axis_title,
    )
    probe_line = generate_probe_line(
        main_chart=main_chart, probe_line_config=probe_line_config
    )
    error_line = generate_error_line(
        main_chart=main_chart, error_line_config=error_line_config
    )
    model_line = generate_model_line(
        main_chart=main_chart, model_line_config=model_line_config
    )
    main_chart = probe_line + error_line + model_line
    if index_selection:
        main_chart = main_chart.add_params(index_selection)

    for estimate_selection in estimates_selections:
        main_chart = main_chart.add_params(estimate_selection)

    main_chart = configure_style(chart=main_chart, chart_style=chart_style)

    if save_path:
        save_html(
            obj=main_chart,
            filename=save_path,
        )

    return main_chart