draw_graphs.py

import plotly.express as px
import sys
import geopandas as gpd
import plotly.graph_objects as go

sys.path.append("./")
import config
import process_data


def get_geo_data():

    geo_df = gpd.read_file("data/NHS_England_Regions_April_2021_EN_BUC_2022.geojson")
    geo_df = geo_df.to_crs(epsg="4326")
    geo_df = geo_df[["NHSER21NM", "geometry"]].set_index("NHSER21NM")

    return geo_df


def set_up_map(df, geo_df, org_level, dimension, selectedpoints):
    fig = px.choropleth_mapbox(
        df,
        geojson=geo_df,
        locations="Org_Name",
        color=config.measure_dict[org_level][dimension]["rate_col"],
        color_continuous_scale=config.nhs_colours,
        mapbox_style="carto-positron",
        center={"lat": 53, "lon": -2},
        zoom=5.2,
    )

    if org_level == "Provider":
        fig = add_scatter_points(fig, df, org_level, dimension)

    fig.update_traces(
        customdata=df[
            ["Org_Name", config.measure_dict[org_level][dimension]["rate_col"]]
        ].values,
        hovertemplate=(
            "%{customdata[0]}<br>"
            "%{customdata[1]:.2f}"
            + ("%" if dimension not in config.special_dimensions else "")
            + "<extra></extra>"
        ),
    )

    fig.update_layout(clickmode="event+select")
    fig.update_coloraxes(colorbar={"orientation": "h", "title": ""})
    fig.update_layout(coloraxis_colorbar_y=-0.1)

    if selectedpoints is not None:
        fig.update_traces(selectedpoints=selectedpoints)

    return fig


def add_scatter_points(fig, df, org_level, dimension):
    # For provider level, adds the scatter points
    fig.add_trace(
        go.Scattermapbox(
            lat=df["latitude"],
            lon=df["longitude"],
            mode="markers",
            marker=go.scattermapbox.Marker(
                size=20,
                color=df[config.measure_dict[org_level][dimension]["rate_col"]],
                colorscale=config.nhs_colours,
                colorbar=dict(title="", orientation="h", y=-0.1),
            ),
        )
    )
    return fig


def draw_map(org_level, dimension, year, selectedpoints=None):
    # pulls together necessary functions to draw either region or provider map
    df = process_data.return_data_for_map(
        dimension, org_level, config.measure_dict, year
    )
    geo_df = get_geo_data()
    fig = set_up_map(df, geo_df, org_level, dimension, selectedpoints)

    return fig


def draw_special_bar_chart(dimension, year):

    df = process_data.return_data_for_special_bar_chart(dimension, year)
    fig = px.bar(df, x="Org_Name", y="Rate")
    fig.update_layout(xaxis_title="", yaxis_title="")

    return fig


def draw_bar_chart(org_level, dimension, year, location):
    df_location = process_data.return_data_for_bar_chart(
        dimension, org_level, location, year
    )
    df_all_submitters = process_data.return_data_for_bar_chart(
        dimension, "National", "All Submitters", year
    )

    # Merge together the df with the All Submitters data to get marker data
    df_merged = process_data.merge_total_submitters(df_location, df_all_submitters)

    # Create the bar chart
    fig = px.bar(df_merged, x="Measure", y="Value")

    # Add custom markers for All Submitters
    fig.add_trace(
        go.Scatter(
            x=df_merged["Measure"],
            y=df_merged["All Submitters Value"],
            mode="markers",
            name="All Submitters",
            marker=dict(symbol="cross", size=10, color="red"),
        )
    )

    fig.update_layout(
        legend={"yanchor": "middle", "xanchor": "left", "y": 0.98, "x": 0}
    )
    fig.update_layout(xaxis_title="", yaxis_title="")

    return fig


def draw_time_series(org_level, dimension, location):
    if dimension in config.special_dimensions:
        split_on = "Org_Name"
        value = "Rate"
    else:
        split_on = "Measure"
        value = "Value"

    df_location = process_data.return_data_for_time_series(
        dimension, org_level, location
    )
    # Strip the first characters to get numeric year for sorting
    df_location["numeric_year"] = df_location["year"].str.split("-").str[0].astype(int)

    # Sort by numeric year
    df_location = df_location.sort_values(by="numeric_year")

    # Create custom hover text
    df_location["hover_text"] = df_location.apply(
        lambda row: f"Year: {row['year']}<br>{split_on}: {row[split_on]}<br>Value: {row[value]}",
        axis=1,
    )

    # Create the time series line graph with custom hover text
    fig = px.line(
        df_location, x="year", y=value, color=split_on, hover_data={"hover_text": True}
    )

    # Update hover data to use custom text
    fig.update_traces(hovertemplate="%{customdata}")

    # Add round, black dots at each point in the graph
    for measure in df_location[split_on].unique():
        df_measure = df_location[df_location[split_on] == measure]

        fig.add_trace(
            go.Scatter(
                x=df_measure["year"],
                y=df_measure[value],
                mode="markers",
                name=f"{measure} Dots",
                marker=dict(symbol="circle", size=6, color="black"),
                hoverinfo="skip",
                showlegend=False,
            )
        )

    fig.update_layout(xaxis_title="", yaxis_title="")

    return fig