Session 14 lab: From analysis to a decision dashboard¶

# Course data setup. Run this cell first; you do not need to understand it yet.
# It downloads the Superstore sample from Tableau's website, checks it is the
# exact version this course was written for, and prepares the order-lines table.
import hashlib
import os
import urllib.request

import matplotlib.pyplot as plt
import pandas as pd

SOURCE_URL = "https://public.tableau.com/app/sample-data/sample_-_superstore.xls"
EXPECTED_SHA256 = "5045ed4fec49cae2cc0dfc5ab0d204ba09efa767fe00c996822499e8cc42fbbd"
SOURCE_FILE = os.environ.get("COURSE_XLS_PATH", "sample_-_superstore.xls")

try:
    import xlrd  # noqa: F401  (needed to read .xls files)
except ImportError:
    import subprocess
    import sys

    subprocess.check_call([sys.executable, "-m", "pip", "install", "-q", "xlrd"])

if not os.path.exists(SOURCE_FILE):
    urllib.request.urlretrieve(SOURCE_URL, SOURCE_FILE)

with open(SOURCE_FILE, "rb") as f:
    digest = hashlib.sha256(f.read()).hexdigest()
if digest != EXPECTED_SHA256:
    print("WARNING: Tableau has changed this sample file since the course was written.")
    print("Your numbers will differ from the lesson. Ask your facilitator for the course copy.")


def load_orders():
    """Return one row per order line, orders placed 2023-01-01 to 2025-12-31."""
    raw = pd.read_excel(SOURCE_FILE, sheet_name="Orders")
    returns = pd.read_excel(SOURCE_FILE, sheet_name="Returns")

    # Orders with more than one Customer ID (a known data issue), found before filtering.
    ids_per_order = raw.groupby("Order ID")["Customer ID"].nunique()
    conflict_orders = set(ids_per_order[ids_per_order > 1].index)

    in_scope = (raw["Order Date"] >= "2023-01-01") & (raw["Order Date"] < "2026-01-01")
    df = raw.loc[in_scope].copy()

    df = df.rename(columns={
        "Row ID": "row_id", "Order ID": "order_id", "Order Date": "order_date",
        "Ship Date": "ship_date", "Ship Mode": "ship_mode", "Customer ID": "customer_id",
        "Segment": "segment", "Country/Region": "country_region", "Region": "region",
        "Category": "category", "Sub-Category": "sub_category", "Product ID": "product_id",
        "Product Name": "product_name", "Sales": "sales", "Quantity": "quantity",
        "Discount": "discount", "Profit": "profit",
    })
    df["returned_order_flag"] = df["order_id"].isin(set(returns["Order ID"])).astype(int)
    df["customer_order_conflict"] = df["order_id"].isin(conflict_orders).astype(int)
    df["customer_analysis_eligible"] = 1 - df["customer_order_conflict"]
    df["order_to_ship_days"] = (df["ship_date"] - df["order_date"]).dt.days
    df["order_month"] = df["order_date"].dt.to_period("M").dt.to_timestamp()

    columns = [
        "row_id", "order_id", "order_date", "ship_date", "ship_mode", "customer_id",
        "segment", "country_region", "region", "category", "sub_category", "product_id",
        "product_name", "sales", "quantity", "discount", "profit", "returned_order_flag",
        "customer_analysis_eligible", "customer_order_conflict", "order_to_ship_days",
        "order_month",
    ]
    return df[columns].sort_values("row_id").reset_index(drop=True)


pd.set_option("display.float_format", "{:,.2f}".format)
plt.rcParams.update({
    "figure.figsize": (8, 4),
    "axes.spines.top": False,
    "axes.spines.right": False,
    "axes.grid": True,
    "axes.axisbelow": True,
    "grid.color": "#e6e5e1",
    "axes.prop_cycle": plt.cycler(color=["#2a78d6", "#eb6834", "#1baf7a"]),
    "axes.titleweight": "bold",
    "axes.titlelocation": "left",
})
print("Setup complete.")

Setup complete.

# Marketing data setup (synthetic). Run this cell first.
# Downloads the course's synthetic marketing data from GitHub and checks it is the exact version this course uses.
import hashlib
import io
import os
import urllib.request

import pandas as pd

MARKETING_URL = "https://raw.githubusercontent.com/Aayush181509/AnalyticsLecture/main/data/synthetic/marketing-v1/"
MARKETING_SHA256 = {
    "marketing_daily.csv": "1cbfc8e1c91ae72d18364b2d13bdf6db513c4aec28c339ef8b91f87a60dc0a07",
    "ab_test.csv": "41a3662997111b66d7ab1eb899f4b32e26bd64c66284ce881db6b8a7300ad47f",
}
MARKETING_DIR = os.environ.get("COURSE_MARKETING_DIR", "")


def load_marketing(name):
    """Load one marketing-v1 file: 'marketing_daily.csv' or 'ab_test.csv'."""
    if MARKETING_DIR:
        with open(os.path.join(MARKETING_DIR, name), "rb") as f:
            data = f.read()
    else:
        data = urllib.request.urlopen(MARKETING_URL + name).read()
    if hashlib.sha256(data).hexdigest() != MARKETING_SHA256[name]:
        print("WARNING: this file differs from the course version. Ask your facilitator.")
    return pd.read_csv(io.BytesIO(data), parse_dates=["date"])


print("Marketing setup complete.")

Marketing setup complete.

import re
import textwrap

import matplotlib.dates as mdates
import numpy as np
from matplotlib.ticker import StrMethodFormatter

THOUSANDS = StrMethodFormatter("{x:,.0f}")             # axis ticks like 20,000
GREY, BLUE, ORANGE, AQUA = "#b8b6b0", "#2a78d6", "#eb6834", "#1baf7a"

df = load_orders()
df["year"] = df["order_date"].dt.year
marketing = load_marketing("marketing_daily.csv")
print("Retail order lines:", len(df), "| orders:", df["order_id"].nunique())
print("Marketing rows:", len(marketing), "| campaigns:", marketing["campaign_id"].nunique(),
      "| days:", marketing["date"].nunique())

Retail order lines: 6815 | orders: 3388
Marketing rows: 2555 | campaigns: 7 | days: 365

d25 = df[df["year"] == 2025]
sales_sub = d25.groupby("sub_category")["sales"].sum()


def sales_chart(remove_decoration=False, sort_horizontal=False, zero_baseline=False,
                grey_plus_one=False, direct_labels=False, focus=None, title="Sales by Sub-Category"):
    """2025 Sales by sub-category, with optional decluttering steps."""
    values = sales_sub.sort_values() if sort_horizontal else sales_sub
    if grey_plus_one:
        colours = [ORANGE if s == focus else GREY for s in values.index]
    elif remove_decoration:
        colours = BLUE
    else:
        colours = plt.cm.tab20(np.linspace(0, 1, len(values)))
    style = {} if remove_decoration else {"edgecolor": "black", "hatch": "//"}

    fig, ax = plt.subplots(figsize=(9, 5))
    if sort_horizontal:
        bars = ax.barh(values.index, values, color=colours, **style)
        value_axis, set_limits = ax.xaxis, ax.set_xlim
    else:
        bars = ax.bar(values.index, values, color=colours, **style)
        ax.tick_params(axis="x", rotation=90)
        value_axis, set_limits = ax.yaxis, ax.set_ylim

    if remove_decoration:
        ax.grid(axis="y" if sort_horizontal else "x", visible=False)
        unit = "Sales, 2025 (source monetary units)"
    else:
        ax.set_facecolor("#d9d9d9")
        ax.grid(True, color="black", linewidth=0.8)
        for spine in ax.spines.values():
            spine.set_visible(True)
        ax.legend(bars, values.index, fontsize=6, ncol=2)
        unit = "Sales"

    if zero_baseline:
        set_limits(0, values.max() * 1.2)
    else:
        set_limits(values.quantile(0.2), values.max() * 1.05)      # truncated: small bars vanish

    if direct_labels:
        for i, v in enumerate(values):
            if sort_horizontal:
                ax.text(v, i, f" {v:,.0f}", va="center", fontsize=8)
            else:
                ax.text(i, v, f"{v:,.0f}", ha="center", va="bottom", fontsize=7, rotation=90)
        value_axis.set_ticks([])
    else:
        value_axis.set_major_formatter(THOUSANDS)

    (ax.set_xlabel if sort_horizontal else ax.set_ylabel)(unit)
    ax.set_title(title)
    plt.show()


sales_chart()

# 1. Set each switch to True, ONE at a time, and run the cell. Write down what each step improved.
# 2. Change focus to the sub-category your message is about (for example the largest, or one you are worried about).
# 3. Replace the title with a message title that is true for the chart, for example naming the leader.
sales_chart(
    remove_decoration=False,
    sort_horizontal=False,
    zero_baseline=False,
    grey_plus_one=False,
    direct_labels=False,
    focus="Chairs",
    title="Sales by Sub-Category",
)

monthly = df.groupby("order_month")["sales"].sum()
ratio_by_category = df.groupby("category")["profit"].sum() / df.groupby("category")["sales"].sum() * 100
ratio_by_category = ratio_by_category.sort_values()
ship_days = df.drop_duplicates("order_id").groupby("ship_mode")["order_to_ship_days"].median().sort_values()


def three_charts(titles):
    """Draw the three charts with the given titles (wrapped to fit)."""
    fig, axes = plt.subplots(1, 3, figsize=(13, 3.8))
    axes[0].plot(monthly.index, monthly, color=BLUE)
    axes[0].set_ylabel("Sales (source monetary units)")
    axes[0].yaxis.set_major_formatter(THOUSANDS)
    axes[0].xaxis.set_major_locator(mdates.YearLocator())
    axes[0].xaxis.set_major_formatter(mdates.DateFormatter("%Y"))
    axes[0].set_xlabel("Order month")
    axes[1].barh(ratio_by_category.index, ratio_by_category, color=GREY)
    axes[1].set_xlabel("Profit ratio, 2023–2025 (%)")
    axes[2].barh(ship_days.index, ship_days, color=GREY)
    axes[2].set_xlabel("Median days from order to shipping")
    for ax, title in zip(axes, titles):
        ax.set_title(textwrap.fill(title, 38), fontsize=10)
    fig.tight_layout()
    plt.show()


print(ratio_by_category.round(1).to_string())
print(ship_days.to_string())

category
Furniture          3.10
Technology        16.90
Office Supplies   17.60
ship_mode
Same Day         0.00
First Class      2.00
Second Class     3.00
Standard Class   5.00

# 1. Run the cell to see the three charts with descriptive titles.
# 2. Replace each title with a MESSAGE title: what should the reader notice? Keep each true and checkable.
# 3. Optional: change the colour of the one bar your title is about to ORANGE in the function above.
titles = [
    "Monthly Sales, 2023–2025",
    "Profit ratio by category",
    "Median days from order to shipping by ship mode",
]
three_charts(titles)

ALL_MONTHS = pd.date_range("2023-01-01", "2025-12-01", freq="MS")


def three_panel_dashboard(data, scope):
    """Monthly trend, segment Sales and returns-list share by region, with a scope note."""
    trend = data.groupby("order_month")["sales"].sum().reindex(ALL_MONTHS, fill_value=0)
    this_year = data[data["year"] == 2025]
    growth = this_year["sales"].sum() / data.loc[data["year"] == 2024, "sales"].sum() - 1
    segments = this_year.groupby("segment")["sales"].sum().sort_values()
    regions = (this_year[this_year["returned_order_flag"] == 1].groupby("region")["order_id"].nunique()
               / this_year.groupby("region")["order_id"].nunique()).fillna(0).sort_values()

    fig = plt.figure(figsize=(12, 8))
    gs = fig.add_gridspec(3, 2, height_ratios=[2, 2, 0.25], hspace=0.55, wspace=0.35)

    ax = fig.add_subplot(gs[0, :])
    ax.plot(trend.index, trend, color=BLUE, linewidth=2)
    ax.set_ylim(0, None)
    ax.yaxis.set_major_formatter(THOUSANDS)
    ax.set_title(f"Sales {growth:+.0%} in 2025 against 2024, with a peak towards the end of every year")
    ax.set_ylabel("Sales (source monetary units)")
    ax.set_xlabel("Order month")

    ax = fig.add_subplot(gs[1, 0])
    ax.barh(segments.index, segments, color=GREY)
    ax.xaxis.set_major_formatter(THOUSANDS)
    ax.grid(axis="y", visible=False)
    ax.set_title(f"{segments.idxmax()} is the largest segment in 2025")
    ax.set_xlabel("Sales, 2025 (source monetary units)")

    ax = fig.add_subplot(gs[1, 1])
    ax.barh(regions.index, regions * 100, color=[ORANGE if r == regions.idxmax() else GREY for r in regions.index])
    for i, v in enumerate(regions):
        ax.text(v * 100, i, f" {v:.1%}", va="center", fontsize=9)
    ax.set_xlim(0, regions.max() * 100 * 1.3)
    ax.grid(axis="y", visible=False)
    ax.set_title(f"{regions.idxmax()} has the highest share of orders on the returns list")
    ax.set_xlabel("% of 2025 orders on the returns list")

    note = fig.add_subplot(gs[2, :])
    note.axis("off")
    note.text(0, 0.5, scope, fontsize=8, color="#555555", va="center", transform=note.transAxes)
    fig.subplots_adjust(left=0.1, right=0.97, top=0.95, bottom=0.03)
    plt.show()


scope_note = ("Scope: Superstore orders placed Jan 2023–Dec 2025 (fictitious sample published by Tableau); "
              "all regions. Sales in source monetary units.\nReturns list = which orders appear on it, "
              "with no dates, amounts or reasons.")
three_panel_dashboard(df, scope_note)

# 1. Change scope_note to "" (empty) and run. Which questions about the page can a reader no longer answer?
# 2. Restore the note. Then redraw the page for one segment: change "All" to "Consumer" (or another segment)
#    and add the segment to the scope note. Is every title still true? Rewrite any that is not.
segment = "All"
data = df if segment == "All" else df[df["segment"] == segment]
three_panel_dashboard(data, scope_note)

MONTH_NAMES = ["Jan", "Feb", "Mar", "Apr", "May", "Jun", "Jul", "Aug", "Sep", "Oct", "Nov", "Dec"]


def case_inputs(case):
    """Everything the dashboard needs for one case: tiles, three views, default titles, scope and facts."""
    if case == "retail":
        d24, d25 = df[df["year"] == 2024], df[df["year"] == 2025]
        sales24, sales25 = d24["sales"].sum(), d25["sales"].sum()
        orders25 = d25["order_id"].nunique()
        listed25 = d25.loc[d25["returned_order_flag"] == 1, "order_id"].nunique()
        seg = (d25.groupby("segment")["sales"].sum() / d24.groupby("segment")["sales"].sum() - 1) * 100
        two = df[df["year"].isin([2024, 2025])].assign(month=lambda t: t["order_date"].dt.month)
        trend = two.pivot_table(index="month", columns="year", values="sales", aggfunc="sum")
        trend.columns = trend.columns.astype(str)
        reg = (d25[d25["returned_order_flag"] == 1].groupby("region")["order_id"].nunique()
               / d25.groupby("region")["order_id"].nunique()).fillna(0) * 100
        return {
            "tiles": [("Sales, 2025", f"{sales25:,.0f}", f"{sales25 / sales24 - 1:+.0%} vs 2024"),
                      ("Orders, 2025", f"{orders25:,}", f"{orders25 / d24['order_id'].nunique() - 1:+.0%} vs 2024"),
                      ("Profit ratio, 2025", f"{d25['profit'].sum() / sales25:.1%}",
                       f"{d24['profit'].sum() / sales24:.1%} in 2024"),
                      ("Orders on returns list, 2025", f"{listed25 / orders25:.1%}", f"{listed25} of {orders25:,} orders")],
            "compare": seg.sort_values(), "compare_label": "Sales growth, 2025 vs 2024 (%)", "compare_fmt": "{:+.0f}%",
            "trend": trend, "trend_label": "Sales (source monetary units)",
            "detail": reg.sort_values(), "detail_label": "% of 2025 orders on the returns list", "detail_fmt": "{:.1f}%",
            "titles": {"headline": "Retail dashboard, 2025",
                       "compare": "Sales growth by segment",
                       "trend": "Monthly Sales, 2024 and 2025",
                       "detail": "Share of orders on the returns list by region"},
            "highlight": {"compare": "Corporate", "detail": "West"},
            "scope": ("Scope: Superstore orders placed Jan 2024–Dec 2025 (fictitious sample published by Tableau). "
                      "Sales in source monetary units. Profit ratio = Profit ÷ Sales.\nReturns list = which orders "
                      "appear on it; no dates, amounts or reasons. Patterns describe what happened, not why."),
            "facts": [f"Sales 2025: {sales25:,.0f} ({sales25 / sales24 - 1:+.1%} vs 2024)",
                      f"Orders 2025: {orders25:,}",
                      "Sales growth by segment (%): " + ", ".join(f"{k} {v:+.1f}" for k, v in seg.items()),
                      f"Share of 2025 orders on the returns list: {listed25 / orders25:.1%}",
                      "By region (%): " + ", ".join(f"{k} {v:.1f}" for k, v in reg.sort_values().items()),
                      f"Months of 2025 above the same month of 2024: {int((trend['2025'] > trend['2024']).sum())} of 12"],
        }
    if case == "marketing":
        m = marketing
        ch = m.groupby("channel")[["spend", "conversions", "attributed_revenue"]].sum()
        roas = (ch["attributed_revenue"] / ch["spend"]).where(ch["spend"] > 0)          # ratio of sums
        cpa = (ch["spend"] / ch["conversions"]).where(ch["conversions"] > 0)
        by_month = m.groupby(m["date"].dt.month)[["spend", "attributed_revenue"]].sum()
        by_month.columns = ["Spend", "Attributed revenue"]
        spend, revenue, conversions = ch["spend"].sum(), ch["attributed_revenue"].sum(), ch["conversions"].sum()
        return {
            "tiles": [("Spend, 2025", f"{spend:,.0f}", "currency units"),
                      ("Attributed revenue, 2025", f"{revenue:,.0f}", "currency units"),
                      ("Return on ad spend", f"{revenue / spend:.2f}", "revenue per unit of spend"),
                      ("Cost per acquisition", f"{spend / conversions:,.2f}", "spend per conversion")],
            "compare": roas.sort_values(), "compare_label": "Attributed revenue per unit of spend (ROAS)",
            "compare_fmt": "{:.2f}",
            "trend": by_month, "trend_label": "Currency units",
            "detail": cpa.sort_values(), "detail_label": "Spend per conversion (CPA, currency units)",
            "detail_fmt": "{:,.0f}",
            "titles": {"headline": "Marketing dashboard, 2025",
                       "compare": "Return on ad spend by channel",
                       "trend": "Monthly spend and attributed revenue",
                       "detail": "Cost per acquisition by channel"},
            "highlight": {"compare": "Search", "detail": "Video"},
            "scope": (f"Scope: synthetic marketing-v1 data, {m['campaign_id'].nunique()} paid campaigns, every day of 2025. "
                      "Currency units. Rates are ratios of sums.\nAttributed revenue is a synthetic single-campaign "
                      "allocation, not proof of extra sales caused. Email has very little media spend."),
            "facts": [f"Spend 2025: {spend:,.0f} | attributed revenue: {revenue:,.0f} | ROAS: {revenue / spend:.2f}",
                      "ROAS by channel: " + ", ".join(f"{k} {v:.2f}" for k, v in roas.sort_values().items()),
                      "CPA by channel: " + ", ".join(f"{k} {v:,.1f}" for k, v in cpa.sort_values().items()),
                      "Share of spend by channel (%): " + ", ".join(
                          f"{k} {v:.1f}" for k, v in (ch["spend"] / spend * 100).sort_values().items()),
                      f"Highest-revenue month: {MONTH_NAMES[by_month['Attributed revenue'].idxmax() - 1]}"],
        }
    raise ValueError('case must be "retail" or "marketing"')


def kpi_tile(ax, label, value, note):
    """A text-only KPI tile: label, big number, comparison or unit."""
    ax.axis("off")
    ax.add_patch(plt.Rectangle((0, 0), 1, 1, transform=ax.transAxes, color="#f3f2ef", zorder=0))
    ax.text(0.06, 0.78, label, transform=ax.transAxes, fontsize=9, color="#555555", va="center")
    ax.text(0.06, 0.46, value, transform=ax.transAxes, fontsize=17, fontweight="bold", va="center")
    ax.text(0.06, 0.16, note, transform=ax.transAxes, fontsize=9, color="#555555", va="center")


def ranked_bars(ax, values, highlight, fmt, label, title):
    """Sorted horizontal bars: grey context, one orange highlight, values written on the bars."""
    ax.barh(values.index, values, color=[ORANGE if k == highlight else GREY for k in values.index])
    for i, v in enumerate(values):
        ax.text(v, i, " " + fmt.format(v), va="center", ha="left" if v >= 0 else "right", fontsize=9)
    low = min(0, values.min() * 1.4)
    ax.set_xlim(low, max(values.max(), 0) * 1.35)
    ax.axvline(0, color="black", linewidth=0.8)
    ax.set_xticks([])
    ax.grid(False)
    ax.spines["bottom"].set_visible(False)
    ax.set_xlabel(label)
    ax.set_title(textwrap.fill(title, 48))


def decision_dashboard(inputs, titles, highlight, show=True):
    """Draw the one-page dashboard: KPI tiles, a trend, two ranked comparisons and the scope note."""
    fig = plt.figure(figsize=(12, 10))
    gs = fig.add_gridspec(4, 4, height_ratios=[0.75, 2.1, 2.1, 0.3], hspace=0.5, wspace=0.45)
    for i, (label, value, note) in enumerate(inputs["tiles"]):
        kpi_tile(fig.add_subplot(gs[0, i]), label, value, note)

    ax = fig.add_subplot(gs[1, :])
    trend = inputs["trend"]
    for j, column in enumerate(trend.columns):
        last = j == len(trend.columns) - 1
        ax.plot(trend.index, trend[column], color=BLUE if last else GREY, linewidth=2.5 if last else 2,
                marker="o", markersize=4)
        ax.text(12.15, trend[column].iloc[-1], column, color=BLUE if last else "#777777", va="center",
                fontweight="bold" if last else "normal")
    ax.set_xticks(range(1, 13), MONTH_NAMES)
    ax.set_xlim(0.7, 13.2)
    ax.set_ylim(0, None)
    ax.yaxis.set_major_formatter(THOUSANDS)
    ax.set_ylabel(inputs["trend_label"])
    ax.set_title(titles["trend"])

    ranked_bars(fig.add_subplot(gs[2, :2]), inputs["compare"], highlight.get("compare"), inputs["compare_fmt"],
                inputs["compare_label"], titles["compare"])
    ranked_bars(fig.add_subplot(gs[2, 2:]), inputs["detail"], highlight.get("detail"), inputs["detail_fmt"],
                inputs["detail_label"], titles["detail"])

    note = fig.add_subplot(gs[3, :])
    note.axis("off")
    note.text(0, 0.5, inputs["scope"], fontsize=8, color="#555555", va="center", transform=note.transAxes)
    fig.suptitle(titles["headline"], x=0.02, y=0.96, ha="left", fontsize=14, fontweight="bold")
    fig.subplots_adjust(left=0.1, right=0.9, top=0.9, bottom=0.02)
    if show:
        plt.show()
    return fig


print("Dashboard functions ready.")

Dashboard functions ready.

# 1. Run the cell as it is: the retail dashboard with DESCRIPTIVE default titles.
# 2. If you prefer, change CASE to "marketing" and run again. Stay with one case for the rest of the lab.
# 3. Replace the four titles with MESSAGE titles that are true for your case (start with the headline).
# 4. Change the highlights to the one bar in each chart that your message is about (or None for no highlight).
CASE = "retail"
inputs = case_inputs(CASE)

titles = dict(inputs["titles"])
# titles["headline"] = "Write your message headline here"
# titles["trend"] = "..."
# titles["compare"] = "..."
# titles["detail"] = "..."

highlight = dict(inputs["highlight"])
# highlight["compare"] = "..."

fig = decision_dashboard(inputs, titles, highlight)

print(f"Fact sheet: {CASE}")
for fact in inputs["facts"]:
    print("-", fact)

Fact sheet: retail
- Sales 2025: 613,934 (+29.8% vs 2024)
- Orders 2025: 1,340
- Sales growth by segment (%): Consumer +12.1, Corporate +60.8, Home Office +39.2
- Share of 2025 orders on the returns list: 5.7%
- By region (%): Central 2.3, East 2.8, South 3.7, West 11.9
- Months of 2025 above the same month of 2024: 11 of 12

# 1. Replace the big idea and the five brief sentences with your own, using numbers from the fact sheet.
# 2. Run the cell and fix anything marked CHECK.
# 3. Read your big idea aloud. Could the commercial lead repeat it after hearing it once?
big_idea = "Write your one-sentence big idea here."
brief = [
    "Recommendation: write what the leadership team should do, and by when.",
    "Evidence: write the first number and what it is compared with.",
    "Evidence: write the second number and what it is compared with.",
    "Limit: write what the data cannot show, or what could change the answer.",
    "Next step: write the check or decision needed next, and who does it.",
]


def sentence_count(text):
    """Count sentences: a full stop, ! or ? followed by a space or the end (so 5.7% is not split)."""
    return len([s for s in re.split(r"[.!?](?:\s|$)", text) if s.strip()])


words = len(big_idea.split())
print("Big idea:", big_idea)
print("  sentences:", sentence_count(big_idea), "| words:", words,
      "| OK" if sentence_count(big_idea) == 1 and words <= 35 else "| CHECK: one sentence, at most about 35 words")
print("Brief sentences:", len(brief), "| OK" if len(brief) == 5 else "| CHECK: exactly five sentences")
for label, sentence in zip(["Evidence 1", "Evidence 2"], brief[1:3]):
    has_number = any(ch.isdigit() for ch in sentence)
    print(f"{label} quotes a number:", "OK" if has_number else "CHECK: add a number from the fact sheet")
print()
print("\n".join(brief))

Big idea: Write your one-sentence big idea here.
  sentences: 1 | words: 6 | OK
Brief sentences: 5 | OK
Evidence 1 quotes a number: CHECK: add a number from the fact sheet
Evidence 2 quotes a number: CHECK: add a number from the fact sheet

Recommendation: write what the leadership team should do, and by when.
Evidence: write the first number and what it is compared with.
Evidence: write the second number and what it is compared with.
Limit: write what the data cannot show, or what could change the answer.
Next step: write the check or decision needed next, and who does it.

# 1. Run the cell. Can you still find the highlighted bar in each chart? How?
# 2. Change the highlight to another bar and run again. Is it still easy to find without colour?
# 3. If not, what would you add (a label, a bold name, an annotation)?
fig = decision_dashboard(inputs, titles, highlight, show=False)
fig.canvas.draw()
rgb = np.asarray(fig.canvas.buffer_rgba())[..., :3] / 255
plt.close(fig)
grey_image = rgb @ np.array([0.299, 0.587, 0.114])     # standard brightness weights

fig, ax = plt.subplots(figsize=(10, 8.5))
ax.imshow(grey_image, cmap="gray", vmin=0, vmax=1)
ax.axis("off")
ax.set_title("Your dashboard in shades of grey")
plt.show()

def risky_dashboard(fix_axis=False, fix_colours=False, fix_rate=False, add_scope=False, full_window=False):
    """A regional dashboard with five switchable problems."""
    d25 = df[df["year"] == 2025]
    region_sales = d25.groupby("region")["sales"].sum()
    orders = d25.groupby("region")["order_id"].nunique()
    listed = d25[d25["returned_order_flag"] == 1].groupby("region")["order_id"].nunique()
    share = (listed / orders).sort_values()
    overall = listed.sum() / orders.sum() if fix_rate else share.mean()
    monthly = df.groupby("order_month")["sales"].sum()
    window = monthly if full_window else monthly.loc["2024-12-01":"2025-02-01"]
    palette = {"Central": BLUE, "East": ORANGE, "South": AQUA, "West": "#8e6bbf"}

    fig = plt.figure(figsize=(12, 6.5))
    gs = fig.add_gridspec(3, 3, height_ratios=[0.75, 2.2, 0.3], hspace=0.5, wspace=0.4)
    kpi_tile(fig.add_subplot(gs[0, 0]), "Orders on returns list, 2025", f"{overall:.1%}",
             "listed orders ÷ all orders" if fix_rate else "average of the region rates")
    if full_window:                       # a fair headline: whole year against whole year
        yearly = df.groupby("year")["sales"].sum()
        kpi_tile(fig.add_subplot(gs[0, 1]), "Sales growth, 2025 vs 2024", f"{yearly[2025] / yearly[2024] - 1:+.0%}",
                 "full calendar years")
    else:                                 # first month shown against last month shown
        kpi_tile(fig.add_subplot(gs[0, 1]), "Sales change in the trend window",
                 f"{window.iloc[-1] / window.iloc[0] - 1:+.0%}", f"{window.index[0]:%b %Y} to {window.index[-1]:%b %Y}")

    ax = fig.add_subplot(gs[1, 0])
    colours = [ORANGE if r == "West" else GREY for r in region_sales.index] if fix_colours else \
        [palette[r] for r in region_sales.index]
    ax.bar(region_sales.index, region_sales, color=colours)
    if fix_axis:
        ax.set_ylim(0, region_sales.max() * 1.1)
        smallest, largest = region_sales.idxmin(), region_sales.idxmax()
        ax.set_title(f"{smallest} sold {region_sales.min() / region_sales.max():.0%} of {largest}'s Sales")
    else:
        ax.set_ylim(90_000, region_sales.max() * 1.02)
        ax.set_title("South sells a fraction of the others")
    ax.yaxis.set_major_formatter(THOUSANDS)
    ax.set_ylabel("Sales, 2025 (source monetary units)")

    ax = fig.add_subplot(gs[1, 1])
    ax.plot(window.index, window, color=BLUE if fix_colours else ORANGE, marker="o")
    ax.set_title("Sales repeat a winter dip every year" if full_window else "Sales are falling fast")
    ax.yaxis.set_major_formatter(THOUSANDS)
    ax.set_ylabel("Sales (source monetary units)")
    ax.set_xticks(pd.date_range("2023-01-01", "2026-01-01", freq="YS") if full_window else window.index)
    ax.xaxis.set_major_formatter(mdates.DateFormatter("%b %Y"))

    ax = fig.add_subplot(gs[1, 2])
    ax.barh(share.index, share * 100, color=[ORANGE if r == "West" else (GREY if fix_colours else BLUE)
                                              for r in share.index])
    ax.set_title("Returns-list share by region")
    ax.set_xlabel("% of 2025 orders on the returns list")

    note = fig.add_subplot(gs[2, :])
    note.axis("off")
    if add_scope:
        note.text(0, 0.5, "Scope: Superstore orders placed in 2025 (trend: Jan 2023–Dec 2025); fictitious sample "
                  "published by Tableau. Sales in source monetary units.\nOrange = West, the region that needs "
                  "attention. Returns list = which orders appear on it; no amounts or reasons.",
                  fontsize=8, color="#555555", va="center", transform=note.transAxes)
    fig.suptitle("Regional review, 2025", x=0.02, y=0.98, ha="left", fontsize=14, fontweight="bold")
    fig.subplots_adjust(left=0.08, right=0.97, top=0.9, bottom=0.03)
    plt.show()


risky_dashboard()

# Turn the switches on one at a time. Record before and after for each.
risky_dashboard(
    fix_axis=False,      # bars start at zero
    fix_colours=False,   # orange means "needs attention" (West) everywhere; grey for context
    fix_rate=False,      # overall share = listed orders ÷ all orders (a ratio of sums)
    add_scope=False,     # period, units, source, meaning of colour and limits
    full_window=False,   # all 36 months instead of three hand-picked ones
)