Py.Cafe

import pandas as pd
from vizro import Vizro
import vizro.models as vm
import vizro.plotly.express as px

# ===============================
# Load main downloads dataset
# ===============================
url = "https://raw.githubusercontent.com/DimedS/kedro-pycafe-data/main/data/02_intermediate/pypi_kedro_downloads.csv"
df = pd.read_csv(url)
df.columns = [c.strip().lower() for c in df.columns]  # project, date, count
df["date"] = pd.to_datetime(df["date"], format="%Y-%m-%d")

# Aggregate by month
df_monthly = (
    df.groupby([df["project"], df["date"].dt.to_period("M")])["count"]
    .sum()
    .reset_index()
)
df_monthly["date"] = df_monthly["date"].dt.to_timestamp()
df_monthly["year"] = df_monthly["date"].dt.year
df_monthly["month"] = df_monthly["date"].dt.month_name().str[:3]

# ===============================
# Page 1: Regular monthly trend + Yearly summaries
# ===============================

# Aggregate yearly downloads
df_yearly = (
    df.groupby([df["project"], df["date"].dt.year])["count"]
    .sum()
    .reset_index()
    .rename(columns={"date": "year", "count": "downloads"})
)

# Line chart: monthly trend
monthly_chart = vm.Graph(
    id="monthly_chart",
    figure=px.line(
        df_monthly,
        x="date",
        y="count",
        color="project",
        markers=True,
        title="Monthly PyPI Downloads (Kedro vs Kedro-Viz)",
        labels={
            "date": "Month",
            "count": "Downloads",
            "project": "Project",
        },
    ),
)
# Convert downloads into millions and round to 1 decimal
df_yearly["downloads_mln"] = (df_yearly["downloads"] / 1_000_000).round(1)

# Bar chart: yearly totals for Kedro
kedro_yearly = df_yearly[df_yearly["project"] == "kedro"]
kedro_bar = vm.Graph(
    id="kedro_yearly",
    figure=px.bar(
        kedro_yearly,
        x="year",
        y="downloads_mln",
        text="downloads_mln",
        title="Kedro Yearly Downloads",
        labels={
            "year": "Year",
            "downloads_mln": "Downloads (M)",
        },
    ).update_traces(texttemplate="%{text:.1f}", textposition="outside"),
)

# Bar chart: yearly totals for Kedro-Viz
viz_yearly = df_yearly[df_yearly["project"] == "kedro-viz"]
viz_bar = vm.Graph(
    id="viz_yearly",
    figure=px.bar(
        viz_yearly,
        x="year",
        y="downloads_mln",
        text="downloads_mln",
        title="Kedro-Viz Yearly Downloads",
        labels={
            "year": "Year",
            "downloads_mln": "Downloads (M)",
        },
    ).update_traces(texttemplate="%{text:.1f}", textposition="outside"),
)

# Put line chart on left, bars stacked on right
page1 = vm.Page(
    title="Monthly & Yearly Downloads",
    layout=vm.Grid(grid=[[0, 1], [0, 2]]),  # left = monthly trend, right top = Kedro bar, right bottom = Viz bar
    components=[monthly_chart, kedro_bar, viz_bar],
)

# ===============================
# Page 2: YoY split charts
# ===============================
# Kedro only
kedro_yoy = df_monthly[df_monthly["project"] == "kedro"]
kedro_chart = vm.Graph(
    id="kedro_yoy_chart",
    figure=px.line(
        kedro_yoy,
        x="month",
        y="count",
        color="year",
        markers=True,
        title="Kedro: Year-over-Year Monthly Downloads",
        labels={
            "month": "Month",
            "count": "Downloads",
            "year": "Year",
        },
    ),
)

# Kedro-Viz only
viz_yoy = df_monthly[df_monthly["project"] == "kedro-viz"]
viz_chart = vm.Graph(
    id="viz_yoy_chart",
    figure=px.line(
        viz_yoy,
        x="month",
        y="count",
        color="year",
        markers=True,
        title="Kedro-Viz: Year-over-Year Monthly Downloads",
        labels={
            "month": "Month",
            "count": "Downloads",
            "year": "Year",
        },
    ),
)

page2 = vm.Page(
    title="YoY Comparison",
    layout=vm.Grid(grid=[[0, 1]]),
    components=[kedro_chart, viz_chart],
)

# ===============================
# Page 3: New Monthly Users + MAU
# ===============================

# New Monthly Kedro Users (split by version)
url_new_kedro_users_monthly = (
    "https://raw.githubusercontent.com/DimedS/kedro-pycafe-data/"
    "refs/heads/main/data/02_intermediate/new_kedro_users_monthly.csv"
)
df_new_users = pd.read_csv(url_new_kedro_users_monthly)
df_new_users.columns = [c.strip().lower() for c in df_new_users.columns]

# Normalise column names in case of hidden spaces or BOM
df_new_users.columns = df_new_users.columns.str.replace(r"[^\w]+", "", regex=True)

# Verify columns
print("Columns:", df_new_users.columns.tolist())

# Parse date
df_new_users["first_year_month"] = pd.to_datetime(df_new_users["first_year_month"], format="%Y-%m")

# Ensure version column exists
if "maxversionprefix" in df_new_users.columns:
    df_new_users.rename(columns={"maxversionprefix": "max_version_prefix"}, inplace=True)

new_users_chart = vm.Graph(
    id="new_users_chart",
    figure=px.bar(
        df_new_users,
        x="first_year_month",
        y="count",
        color="max_version_prefix",
        title="New Monthly Kedro Users by Version",
        labels={
            "first_year_month": "Month",
            "count": "New Users",
            "max_version_prefix": "Kedro Version",
        },
        barmode="stack",
    ),
)


# Kedro MAU (per version)
url_mau_kedro = (
    "https://raw.githubusercontent.com/DimedS/kedro-pycafe-data/"
    "refs/heads/main/data/02_intermediate/mau_kedro.csv"
)
df_mau = pd.read_csv(url_mau_kedro)
df_mau.columns = [c.strip().lower() for c in df_mau.columns]
df_mau["year_month"] = pd.to_datetime(df_mau["year_month"], format="%Y-%m")

mau_chart = vm.Graph(
    id="mau_chart",
    figure=px.bar(
        df_mau,
        x="year_month",
        y="mau",
        color="max_version_prefix",
        title="Kedro Monthly Active Users by Version",
        labels={
            "year_month": "Month",
            "mau": "Monthly Active Users",
            "max_version_prefix": "Kedro Version",
        },
        barmode="stack",
    ),
)

# Put both charts side by side
page3 = vm.Page(
    title="Kedro Telemetry Users",
    layout=vm.Grid(grid=[[0, 1]]),
    components=[new_users_chart, mau_chart],
)

# ===============================
# Page 4: Kedro Commands & Plugins MAU
# ===============================

# --- Commands ---
url_kedro_commands_mau = "https://raw.githubusercontent.com/DimedS/kedro-pycafe-data/refs/heads/main/data/02_intermediate/kedro_commands_mau.csv"
df_cmds = pd.read_csv(url_kedro_commands_mau)
df_cmds.columns = [c.strip().lower() for c in df_cmds.columns]
df_cmds["Year-Month"] = pd.to_datetime(df_cmds["year_month"], format="%Y-%m")
df_cmds = df_cmds.rename(
    columns={
        "first_two_words": "Command",
        "unique_users": "Unique Users"
    }
)

commands_chart = vm.Graph(
    id="commands_chart",
    figure=px.line(
        df_cmds,
        x="Year-Month",
        y="Unique Users",
        color="Command",
        markers=True,
        title="Kedro Commands Usage (MAU by Command)",
        labels={
            "Year-Month": "Month",
            "Unique Users": "Unique Users",
            "Command": "Kedro Command"
        },
    ),
)

# --- Plugins ---
url_kedro_plugins_mau = "https://raw.githubusercontent.com/DimedS/kedro-pycafe-data/refs/heads/main/data/02_intermediate/kedro_plugins_mau.csv"
df_plugins = pd.read_csv(url_kedro_plugins_mau)
df_plugins.columns = [c.strip().lower() for c in df_plugins.columns]
df_plugins["Year-Month"] = pd.to_datetime(df_plugins["year_month"], format="%Y-%m")
df_plugins = df_plugins.rename(
    columns={
        "first_two_words": "Plugin",
        "unique_users": "Unique Users"
    }
)

plugins_chart = vm.Graph(
    id="plugins_chart",
    figure=px.line(
        df_plugins,
        x="Year-Month",
        y="Unique Users",
        color="Plugin",
        markers=True,
        title="Kedro Plugins Usage (MAU by Plugin)",
        labels={
            "Year-Month": "Month",
            "Unique Users": "Unique Users",
            "Plugin": "Kedro Plugin"
        },
    ),
)

# Page with both side by side
page4 = vm.Page(
    title="Kedro Commands & Plugins Telemetry MAU",
    layout=vm.Grid(grid=[[0, 1]]),
    components=[commands_chart, plugins_chart],
)

# ===============================
# Page 5: Downloads by Country (Kedro & Kedro-Viz, 2025)
# ===============================

import pandas as pd
import vizro.models as vm
import vizro.plotly.express as px
from plotly import colors

# --- Load and prepare data ---
url_downloads_by_country = (
    "https://raw.githubusercontent.com/DimedS/kedro-pycafe-data/refs/heads/main/data/02_intermediate/downloads_by_country.csv"
)
df_geo = pd.read_csv(url_downloads_by_country)
df_geo.columns = [c.strip().lower() for c in df_geo.columns]  # country_code, count

# ISO-2 → ISO-3 mapping
iso2_to_iso3 = {
    "US": "USA", "NL": "NLD", "IE": "IRL", "SG": "SGP", "AU": "AUS", "JP": "JPN", "DE": "DEU", "BE": "BEL", "BR": "BRA",
    "CA": "CAN", "GB": "GBR", "PL": "POL", "IN": "IND", "FR": "FRA", "CH": "CHE", "SE": "SWE", "KR": "KOR", "FI": "FIN",
    "RU": "RUS", "CN": "CHN", "ES": "ESP", "IT": "ITA", "CL": "CHL", "IL": "ISR", "PR": "PRI", "HK": "HKG", "TW": "TWN",
    "TH": "THA", "PT": "PRT", "CO": "COL", "AE": "ARE", "MX": "MEX", "AT": "AUT", "VN": "VNM", "ZA": "ZAF", "DK": "DNK",
    "CZ": "CZE", "ID": "IDN", "AR": "ARG", "PE": "PER", "TR": "TUR", "IR": "IRN", "UA": "UKR", "HU": "HUN", "GE": "GEO",
    "PH": "PHL", "LV": "LVA", "NO": "NOR", "GT": "GTM", "SK": "SVK", "PA": "PAN", "MT": "MLT", "CR": "CRI", "KZ": "KAZ",
    "GR": "GRC", "EG": "EGY", "MY": "MYS", "EE": "EST", "LT": "LTU", "NZ": "NZL", "TG": "TGO", "BG": "BGR", "RO": "ROU",
    "LK": "LKA", "TN": "TUN", "EC": "ECU", "SA": "SAU", "SI": "SVN", "RS": "SRB", "PK": "PAK", "UY": "URY", "MA": "MAR",
    "LB": "LBN", "DO": "DOM", "HR": "HRV", "BD": "BGD", "KE": "KEN", "MU": "MUS", "IQ": "IRQ", "DZ": "DZA", "AM": "ARM",
    "NG": "NGA", "IS": "ISL", "BY": "BLR", "AZ": "AZE", "UZ": "UZB", "BA": "BIH", "CY": "CYP", "GH": "GHA", "VE": "VEN",
    "LU": "LUX", "QA": "QAT", "BO": "BOL", "ML": "MLI", "JO": "JOR", "ET": "ETH", "CI": "CIV", "ZW": "ZWE", "NP": "NPL",
    "UG": "UGA", "GI": "GIB", "MG": "MDG", "EU": "EUU", "KH": "KHM", "CD": "COD", "NC": "NCL", "PY": "PRY", "MK": "MKD",
    "CM": "CMR", "HN": "HND", "AL": "ALB", "AO": "AGO", "MM": "MMR", "GF": "GUF", "MW": "MWI", "MD": "MDA", "BH": "BHR",
    "MC": "MCO", "TT": "TTO", "OM": "OMN", "LI": "LIE", "BB": "BRB", "JM": "JAM", "KG": "KGZ", "BW": "BWA", "SV": "SLV",
    "NI": "NIC", "FO": "FRO", "PS": "PSE", "IM": "IMN", "GA": "GAB", "CG": "COG", "NA": "NAM", "SC": "SYC", "SN": "SEN",
    "ME": "MNE", "KW": "KWT"
}
df_geo["iso3"] = df_geo["country_code"].map(iso2_to_iso3)

# --- Compute totals ---
usa_downloads = df_geo.loc[df_geo["country_code"] == "US", "count"].sum()
usa_millions = round(usa_downloads / 1_000_000, 1)

# Exclude USA for map clarity
df_geo = df_geo[df_geo["country_code"] != "US"].dropna(subset=["iso3"])

# Add metrics
df_geo["downloads_k"] = (df_geo["count"] / 1_000).round(1)
total_ex_us = df_geo["count"].sum()
df_geo["percent"] = (df_geo["count"] / total_ex_us * 100).round(1)

# --- Improved colour scale for better contrast ---
contrasted_scale = colors.sequential.Tealgrn

# --- Choropleth map ---
fig_map = px.choropleth(
    df_geo,
    locations="iso3",
    locationmode="ISO-3",
    color="downloads_k",
    color_continuous_scale=contrasted_scale,
    projection="natural earth",
    title=(
        f"Kedro & Kedro-Viz Downloads by Country (2025). "
        f"<sup>Colour scale excludes USA for visibility (USA: {usa_millions}M downloads)</sup>"
    ),
    labels={"downloads_k": "Downloads (k)", "percent": "Share (%)"},
    hover_data=["downloads_k", "percent"],
)

fig_map.update_layout(
    paper_bgcolor="#111111",
    plot_bgcolor="#111111",
    font_color="#F2F2F2",
    geo=dict(
        bgcolor="#D9D9D9", showframe=False, showcountries=True,
        countrycolor="#3C3C3C", landcolor="#1A1A1A", lakecolor="#BFBFBF", coastlinecolor="#444444"
    ),
    coloraxis_colorbar=dict(title="Downloads (k)", tickcolor="#F2F2F2"),
)

# --- Complementary bar chart: top 15 countries ---
top_countries = df_geo.sort_values("count", ascending=False).head(15)

fig_bar = px.bar(
    top_countries,
    x="downloads_k",
    y="iso3",
    orientation="h",
    text="downloads_k",
    # remove color=... so we can control manually below
    title="Top 15 Countries",
    labels={"downloads_k": "Downloads (k)", "iso3": "Country"},
)

# Manually colour the bars (using the same Tealgrn scale)
from plotly import colors
contrasted_scale = colors.sequential.Tealgrn
bar_colors = [
    contrasted_scale[int(i * (len(contrasted_scale) - 1) / (len(top_countries) - 1))]
    for i in range(len(top_countries))
]
fig_bar.update_traces(
    marker_color=bar_colors,
    texttemplate="%{text}k",
    textposition="outside",
)

fig_bar.update_layout(
    paper_bgcolor="#111111",
    plot_bgcolor="#111111",
    font_color="#F2F2F2",
    yaxis=dict(categoryorder="total ascending"),
)

# --- Vizro Page (map 75%, bar chart 25%) ---
page5 = vm.Page(
    title="Downloads by Country",
    layout=vm.Grid(grid=[[0, 0, 0, 0, 1]]),  # map spans 3 cells, bar spans 1 cell
    components=[
        vm.Graph(id="geo_downloads_absolute_map", figure=fig_map),
        vm.Graph(id="top_countries_bar", figure=fig_bar),
    ],
)



# ===============================
# Combine all pages
# ===============================
dashboard = vm.Dashboard(pages=[page1, page2, page3, page4, page5])
Vizro().build(dashboard).run()