Analise Seu Data Lake pelo Cursor IDE com DataSpoc MCP

{
  "mcpServers": {
    "dataspoc-lens": {
      "command": "dataspoc-lens",
      "args": ["mcp"],
      "env": {
        "DATASPOC_BUCKET": "s3://my-company-data"
      }
    }
  }
}

{
  "mcpServers": {
    "dataspoc-lens": {
      "command": "dataspoc-lens",
      "args": ["mcp"],
      "env": {
        "DATASPOC_BUCKET": "gs://my-company-data"
      }
    }
  }
}

What tables are available in my data lake?

Your data lake contains 8 tables:

raw_postgres_orders        - 245,000 rows
raw_postgres_customers     - 18,200 rows
raw_stripe_payments        - 892,000 rows
raw_hubspot_contacts       - 34,500 rows
curated_sales              - 245,000 rows
curated_customer_360       - 18,200 rows
gold_revenue_by_region     - 48 rows
gold_monthly_metrics       - 36 rows

What columns does curated_sales have? Show me sample data.

curated_sales schema:
  order_id        VARCHAR
  customer_id     VARCHAR
  product         VARCHAR
  category        VARCHAR
  amount          DOUBLE
  quantity        INTEGER
  region          VARCHAR
  sale_date       DATE
  payment_method  VARCHAR

Sample data (5 rows):
| order_id | customer_id | product        | amount  | region         | sale_date  |
|----------|-------------|----------------|---------|----------------|------------|
| ORD-001  | CUST-042    | Widget Pro     | 299.99  | North America  | 2026-03-15 |
| ORD-002  | CUST-108    | Widget Basic   | 49.99   | Europe         | 2026-03-15 |
| ORD-003  | CUST-042    | Gadget Plus    | 599.00  | North America  | 2026-03-16 |

Show me revenue trends by month for the last 6 months, broken down by region.

SELECT
    DATE_TRUNC('month', sale_date) AS month,
    region,
    SUM(amount) AS revenue,
    COUNT(*) AS orders,
    ROUND(AVG(amount), 2) AS avg_order_value
FROM curated_sales
WHERE sale_date >= '2025-10-01'
GROUP BY month, region
ORDER BY month, revenue DESC

Write a Python script that analyzes customer cohorts by signup month
and tracks their monthly spending over time. Use dataspoc_lens to query.

"""Customer cohort analysis using DataSpoc Lens."""
import pandas as pd
import matplotlib.pyplot as plt
from dataspoc_lens import LensClient

lens = LensClient()

# Get customer cohorts (signup month) with their spending over time
df = lens.query("""
    WITH cohorts AS (
        SELECT
            customer_id,
            DATE_TRUNC('month', MIN(sale_date)) AS cohort_month
        FROM curated_sales
        GROUP BY customer_id
    ),
    monthly_spend AS (
        SELECT
            s.customer_id,
            DATE_TRUNC('month', s.sale_date) AS spend_month,
            SUM(s.amount) AS monthly_total
        FROM curated_sales s
        GROUP BY s.customer_id, spend_month
    )
    SELECT
        c.cohort_month,
        m.spend_month,
        DATEDIFF('month', c.cohort_month, m.spend_month) AS months_since_signup,
        COUNT(DISTINCT m.customer_id) AS active_customers,
        SUM(m.monthly_total) AS total_revenue,
        AVG(m.monthly_total) AS avg_spend_per_customer
    FROM cohorts c
    JOIN monthly_spend m ON c.customer_id = m.customer_id
    WHERE c.cohort_month >= '2025-07-01'
    GROUP BY c.cohort_month, m.spend_month, months_since_signup
    ORDER BY c.cohort_month, months_since_signup
""")

# Pivot for cohort retention chart
pivot = df.pivot_table(
    index="cohort_month",
    columns="months_since_signup",
    values="active_customers",
    aggfunc="sum"
)

# Calculate retention rates
retention = pivot.div(pivot[0], axis=0) * 100

# Plot
fig, ax = plt.subplots(figsize=(12, 8))
import seaborn as sns
sns.heatmap(
    retention,
    annot=True,
    fmt=".0f",
    cmap="YlOrRd_r",
    ax=ax,
    vmin=0,
    vmax=100,
)
ax.set_title("Customer Cohort Retention (%)")
ax.set_xlabel("Months Since Signup")
ax.set_ylabel("Cohort Month")
plt.tight_layout()
plt.savefig("cohort_retention.png", dpi=150)
plt.show()
print("Chart saved to cohort_retention.png")

Create a bar chart showing the top 10 products by revenue this quarter.

from dataspoc_lens import LensClient
import matplotlib.pyplot as plt

lens = LensClient()

df = lens.query("""
    SELECT product, SUM(amount) AS revenue
    FROM curated_sales
    WHERE sale_date >= '2026-01-01'
    GROUP BY product
    ORDER BY revenue DESC
    LIMIT 10
""")

fig, ax = plt.subplots(figsize=(10, 6))
ax.barh(df["product"][::-1], df["revenue"][::-1])
ax.set_xlabel("Revenue ($)")
ax.set_title("Top 10 Products by Revenue — Q1 2026")

for i, (val, name) in enumerate(zip(df["revenue"][::-1], df["product"][::-1])):
    ax.text(val + 1000, i, f"${val:,.0f}", va="center", fontsize=9)

plt.tight_layout()
plt.savefig("top_products.png", dpi=150)
plt.show()

Write a weekly metrics report for the last 7 days. Include revenue,
order count, top products, and regional breakdown. Format as markdown.

# Weekly Metrics Report: April 8-14, 2026

## Summary
- **Total Revenue**: $987,000 (+8.3% vs prior week)
- **Orders**: 3,420 (+5.1% vs prior week)
- **Average Order Value**: $288.60 (+3.0% vs prior week)

## Daily Revenue
| Date | Revenue | Orders |
|------|---------|--------|
| Apr 8  | $142,000 | 490 |
| Apr 9  | $138,000 | 478 |
| Apr 10 | $145,000 | 502 |
| Apr 11 | $151,000 | 523 |
| Apr 12 | $133,000 | 461 |
| Apr 13 | $128,000 | 443 |
| Apr 14 | $150,000 | 523 |

## Top 5 Products
1. Widget Pro — $124,000 (12.6%)
2. Gadget Plus — $98,000 (9.9%)
...

What tables are available? Show me the schema of the largest table.

Revenue by product category for March 2026 only, excluding returns.

Compare this month vs last month for all key metrics.
Show percentage changes.

Write a function I can reuse to generate monthly reports.
Take the month as a parameter.

First show me which customers churned last month.
Then analyze what they had in common (plan, region, usage).
Then suggest retention strategies based on the patterns.