Py.Cafe

import pandas as pd
import vizro.plotly.express as px
from vizro import Vizro
import vizro.models as vm
import vizro.tables as vt # Import vizro.tables

# --- 1. Data Loading and Preprocessing ---

# The names of the files uploaded from the Excel sheets have been updated
# New files are:
SHOWS_FILE = "25_H1 Netflix Report_Shows.csv"
MOVIES_FILE = "25_H1 Netflix Report_Movies.csv"

# Load dataframes
try:
    # Adding thousands separator and decimal handling for robust loading of data like '313,000,000'
    df_shows = pd.read_csv(SHOWS_FILE, thousands=',')
    df_movies = pd.read_csv(MOVIES_FILE, thousands=',')
except FileNotFoundError:
    print("Error: One or both data files were not found. Please ensure the file paths are correct.")
    # Create empty dataframes to prevent crash in the vizro part
    df_shows = pd.DataFrame(columns=['Title', 'Available Globally?', 'Release Date', 'Hours Viewed', 'Runtime', 'Views'])
    df_movies = pd.DataFrame(columns=['Title', 'Available Globally?', 'Release Date', 'Hours Viewed', 'Runtime', 'Views'])

def preprocess_data(df):
    """Cleans and transforms raw data for visualization."""
    # Convert 'Release Date' to datetime, coercing errors (e.g., if a date is missing)
    df['Release Date'] = pd.to_datetime(df['Release Date'], errors='coerce')
    df['Release Month'] = df['Release Date'].dt.to_period('M').astype(str).fillna('TBD')

    # Convert H:MM runtime string to total hours (float)
    def runtime_to_hours(runtime_str):
        if pd.isna(runtime_str):
            return 0
        try:
            # Handle cases where runtime might be a single number (e.g., if it was parsed as int/float)
            if isinstance(runtime_str, (int, float)):
                return runtime_str
            h, m = map(int, str(runtime_str).split(':'))
            return round(h + m / 60, 2)
        except:
            return 0

    df['Runtime (Hours)'] = df['Runtime'].apply(runtime_to_hours)

    # Standardize column names for easier use in Vizro/Plotly
    df.rename(columns={
        "Available Globally?": "Available_Globally",
        "Hours Viewed": "Hours_Viewed",
    }, inplace=True)

    # Ensure Hours_Viewed and Views are numeric (needed especially after using thousands=',')
    for col in ['Hours_Viewed', 'Views']:
        if col in df.columns:
            df[col] = pd.to_numeric(df[col], errors='coerce').fillna(0)

    return df

df_shows = preprocess_data(df_shows)
df_movies = preprocess_data(df_movies)

# Prepare data for specific charts (Top N lists are often best calculated pre-chart)
shows_top_10 = df_shows.sort_values("Hours_Viewed", ascending=False).head(10).reset_index(drop=True)
movies_top_10 = df_movies.sort_values("Views", ascending=False).head(10).reset_index(drop=True)

# --- 2. Vizro Components (Charts and Tables) ---

# --- Shows Components ---
shows_top_hours_chart = vm.Graph(
    title="Top 10 Shows by Hours Viewed",
    # Added Header and Footer for context, as seen in the demo app
    header="Shows are ranked by their total **Hours Viewed** in the first half of 2025. This metric is the primary indicator of content consumption.",
    footer="Data reveals 'Adolescence: Limited Series' and 'Squid Game: Season 2' were the top performers, contributing significantly to overall viewing hours.",
    figure=px.bar(
        shows_top_10,
        x="Title",
        y="Hours_Viewed",
        color="Available_Globally",
        template="vizro_dark",
        labels={"Hours_Viewed": "Hours Viewed (Millions)", "Available_Globally": "Global Release"}
    )
)

shows_month_hours_chart = vm.Graph(
    title="Total Hours Viewed by Release Month",
    # Added Header and Footer for context, as seen in the demo app
    header="Analysis of total hours viewed grouped by the original content **release month** (H1 2025).",
    footer="Viewing hours show high peaks in January and March, suggesting successful rollouts of major titles during these periods.",
    figure=px.bar(
        df_shows.groupby("Release Month", as_index=False)["Hours_Viewed"].sum().sort_values("Release Month"),
        x="Release Month",
        y="Hours_Viewed",
        template="vizro_dark",
        labels={"Hours_Viewed": "Total Hours Viewed (Millions)"}
    )
)

shows_table = vm.Table(
    id="shows_data_table",
    title="All 2025 H1 Shows Data",
    figure=vt.dash_data_table(data_frame=df_shows[['Title', 'Release Date', 'Available_Globally', 'Hours_Viewed', 'Views', 'Runtime (Hours)']])
)

# --- Movies Components ---
movies_top_views_chart = vm.Graph(
    title="Top 10 Movies by Views",
    # Added Header and Footer for context, as seen in the demo app
    header="Movies are ranked by their total **Views** in the first half of 2025. This indicates overall audience reach and popularity.",
    footer="The movie 'Back in Action' dominated H1 in terms of views, signaling a high-demand title with strong initial audience pull.",
    figure=px.bar(
        movies_top_10,
        x="Title",
        y="Views",
        color="Available_Globally",
        template="vizro_dark",
        labels={"Views": "Views (Millions)", "Available_Globally": "Global Release"}
    )
)

movies_avg_runtime_chart = vm.Graph(
    title="Average Runtime by Release Month",
    # Added Header and Footer for context, as seen in the demo app
    header="The average runtime (in hours) for movies released in each month of H1 2025.",
    footer="The consistency in average runtime suggests a controlled release strategy, with no dramatic variation in film length across the half-year.",
    figure=px.bar(
        df_movies.groupby("Release Month", as_index=False)["Runtime (Hours)"].mean().sort_values("Release Month"),
        x="Release Month",
        y="Runtime (Hours)",
        template="vizro_dark",
        labels={"Runtime (Hours)": "Avg Runtime (Hours)"}
    )
)

movies_table = vm.Table(
    id="movies_data_table",
    title="All 2025 H1 Movies Data",
    figure=vt.dash_data_table(data_frame=df_movies[['Title', 'Release Date', 'Available_Globally', 'Hours_Viewed', 'Views', 'Runtime (Hours)']])
)

# --- 3. Vizro Pages and Dashboard ---

shows_page = vm.Page(
    title="TV Shows Analysis",
    # FIXED: Grid indices must start at 0. [0, 1] for the top row charts, [2, 2] for the table spanning the bottom row.
    layout=vm.Layout(grid=[[0, 1], [2, 2]]),
    components=[
        shows_top_hours_chart,
        shows_month_hours_chart,
        shows_table,
    ],
    controls=[
        # Filters are applied to all components on the page that use the respective columns
        vm.Filter(column="Available_Globally", selector=vm.Dropdown(title="Global Availability")),
        # FIX: Removed explicit selector to allow Vizro to use the correct DatePickerRange component automatically for temporal data.
        vm.Filter(column="Release Date"),
    ],
)

movies_page = vm.Page(
    title="Movies Analysis",
    # FIXED: Grid indices must start at 0. [0, 1] for the top row charts, [2, 2] for the table spanning the bottom row.
    layout=vm.Layout(grid=[[0, 1], [2, 2]]),
    components=[
        movies_top_views_chart,
        movies_avg_runtime_chart,
        movies_table,
    ],
    controls=[
        vm.Filter(column="Available_Globally", selector=vm.Dropdown(title="Global Availability")),
        # FIX: Removed explicit selector to allow Vizro to use the correct DatePickerRange component automatically for temporal data.
        vm.Filter(column="Release Date"),
    ],
)

# Combine pages into the final dashboard
dashboard = vm.Dashboard(
    title="Netflix 2025 H1 Report Dashboard",
    pages=[shows_page, movies_page]
)

# Build and run the dashboard
Vizro().build(dashboard).run()