Data Analytics Reporter

# Remove outliers (IQR method)
Q1, Q3 = df["value"].quantile([0.25, 0.75])
IQR = Q3 - Q1
df = df[df["value"].between(Q1 - 1.5 * IQR, Q3 + 1.5 * IQR)]

# Parse dates with coercion (silently converts unparseable values to NaT)
df["date"] = pd.to_datetime(df["date"], errors="coerce")

# Period-over-period growth
prev_period = df[df["period"] == "Q3"]["revenue"].sum()
curr_period = df[df["period"] == "Q4"]["revenue"].sum()
growth_pct = (curr_period - prev_period) / prev_period * 100

# Conversion rate
conversion_rate = df[df["converted"] == True].shape[0] / df.shape[0] * 100

# Customer segmentation by spend
df["segment"] = pd.cut(
    df["total_spend"],
    bins=[0, 100, 500, float("inf")],
    labels=["Low", "Mid", "High"]
)
segment_summary = df.groupby("segment")["total_spend"].agg(["count", "sum", "mean"])

-- Period-over-period KPI
SELECT
    DATE_TRUNC('month', order_date) AS month,
    SUM(revenue)                    AS total_revenue,
    COUNT(DISTINCT customer_id)     AS unique_customers,
    SUM(revenue) / COUNT(DISTINCT customer_id) AS revenue_per_customer
FROM orders
WHERE order_date >= '2024-01-01'
GROUP BY 1
ORDER BY 1;

-- Cohort retention skeleton
SELECT
    first_order_month,
    DATE_DIFF(order_month, first_order_month, MONTH) AS months_since_first,
    COUNT(DISTINCT customer_id) AS retained_customers
FROM cohort_base
GROUP BY 1, 2
ORDER BY 1, 2;

from scipy import stats

control = df[df["variant"] == "control"]["conversion"]
treatment = df[df["variant"] == "treatment"]["conversion"]

t_stat, p_value = stats.ttest_ind(control, treatment)
print(f"Control mean: {control.mean():.4f}")
print(f"Treatment mean: {treatment.mean():.4f}")
print(f"Lift: {(treatment.mean() - control.mean()) / control.mean() * 100:.2f}%")
print(f"p-value: {p_value:.4f} ({'significant' if p_value < 0.05 else 'not significant'} at 95% CI)")

import matplotlib.pyplot as plt

# Use tight_layout and explicit dpi for all saved charts
plt.tight_layout()
plt.savefig("chart_name.png", dpi=150)

# Always label axes with units and give titles that describe the insight, not just the data
plt.xlabel("Date"); plt.ylabel("Revenue ($)")
plt.title("Weekly Revenue Declined 12% in Q4")  # insight-first, not "Revenue Over Time"

## [Report Title] — [Date Range]

### Key Findings
- [Metric 1]: [Value] ([+/-X%] vs prior period)
- [Metric 2]: [Value] ([+/-X%] vs prior period)
- [Metric 3]: [Value] ([+/-X%] vs prior period)

### Trend Analysis
[2–3 sentences describing the dominant trend and its likely drivers.]

### Segments / Breakdown
| Segment | Value | Share | vs Prior |
|---------|-------|-------|----------|
| ...     | ...   | ...   | ...      |

### Recommendation
[One concrete action supported directly by the data above.]

Dashboard: [Name]

Filters: Date range picker | Region dropdown | Segment multi-select

Row 1 — KPI Tiles (4 across)
  • Total Revenue  |  MoM Growth %  |  Active Users  |  Conversion Rate

Row 2 — Primary Chart
  • Line chart: Daily/Weekly [Metric] over selected date range, with prior-period overlay

Row 3 — Breakdown Charts (2 across)
  • Bar chart: [Metric] by [Dimension]
  • Pie/Donut: Share by [Segment]

Row 4 — Data Table
  • Sortable table: [Dimension] | [Metric A] | [Metric B] | Change %

Refresh cadence: [Daily / Weekly]
Data source: [Table / file name]