Why Telecom Companies Are Hiring More Data Analysts in 2026

-- Call drop rate by site and hour for one day
SELECT site_id,
       strftime('%H', call_start) AS hour,
       COUNT(*) AS calls,
       ROUND(100.0 * SUM(CASE WHEN end_reason = 'DROPPED' THEN 1 ELSE 0 END)
             / COUNT(*), 2) AS drop_rate_pct
FROM call_records
WHERE DATE(call_start) = '2026-11-05'
GROUP BY site_id, hour
HAVING COUNT(*) >= 200
ORDER BY drop_rate_pct DESC
LIMIT 20;

import pandas as pd

kpi = pd.DataFrame({
    "site": ["S1"] * 6,
    "week": [1, 2, 3, 4, 5, 6],
    "drop_rate_pct": [2.8, 3.1, 2.9, 1.6, 1.5, 1.7],
})
fix_week = 4
before = kpi.loc[kpi["week"] < fix_week, "drop_rate_pct"].mean()
after = kpi.loc[kpi["week"] >= fix_week, "drop_rate_pct"].mean()
print(f"before: {before:.2f}%  after: {after:.2f}%")
# Practice data. Check other sites over the same weeks for seasonal effects.

-- Churn rate by plan and tenure band (last quarter)
SELECT plan_name,
       CASE WHEN tenure_months < 6 THEN '0-5 months'
            WHEN tenure_months < 24 THEN '6-23 months'
            ELSE '24+ months' END AS tenure_band,
       COUNT(*) AS subscribers,
       ROUND(100.0 * AVG(CASE WHEN churned THEN 1 ELSE 0 END), 1) AS churn_pct
FROM subscriber_quarter
WHERE quarter = '2026-Q3'
GROUP BY plan_name, tenure_band
ORDER BY churn_pct DESC;

import pandas as pd

subs = pd.DataFrame({
    "user": ["u1", "u2", "u3", "u4"],
    "plan_gb": [42, 42, 84, 84],           # data allowance per cycle
    "used_gb": [41, 12, 30, 83],
})
subs["use_ratio"] = subs["used_gb"] / subs["plan_gb"]
subs["fit"] = pd.cut(subs["use_ratio"], [0, 0.4, 0.9, 10],
                     labels=["under-using", "good fit", "near limit"])
print(subs)
# Near-limit users may want upgrades; under-users may churn to cheaper plans.

import pandas as pd

funnel = pd.DataFrame({
    "channel": ["app"] * 4 + ["retailer"] * 4,
    "step": ["started", "plan_selected", "payment", "success"] * 2,
    "count": [10000, 8200, 7400, 7000, 3000, 2950, 2900, 2880],
})
funnel["from_prev_%"] = (funnel.groupby("channel")["count"]
                          .transform(lambda s: s / s.shift(1) * 100).round(1))
print(funnel)
# Practice numbers. Compare where each channel loses the most users.

-- Top ticket categories and repeat-contact rate last month
SELECT category,
       COUNT(*) AS tickets,
       ROUND(AVG(resolution_hours), 1) AS avg_resolution_hours,
       ROUND(100.0 * AVG(CASE WHEN repeat_within_7d THEN 1 ELSE 0 END), 1)
         AS repeat_contact_pct
FROM support_tickets
WHERE created_at >= '2026-10-01' AND created_at < '2026-11-01'
GROUP BY category
ORDER BY tickets DESC
LIMIT 10;

call_records(call_id, subscriber_id, site_id, call_start, duration_s, end_reason)
subscriber_quarter(subscriber_id, quarter, plan_name, tenure_months, churned)
usage_cycle(subscriber_id, cycle_start, plan_gb, used_gb)
recharges(recharge_id, subscriber_id, channel, method, step, status, ts)
support_tickets(ticket_id, subscriber_id, category, created_at,
                resolution_hours, repeat_within_7d)

Telecom operations dashboard

- Call drop rate by site (top 20) and by hour
- Churn rate by plan and tenure band
- Plan fit: under-using vs near-limit share
- Recharge funnel by channel
- Top support categories with repeat-contact rate

Filters: circle/region, plan, date range

Finding: Short-tenure subscribers on one plan churn noticeably
more than others in this practice dataset.

Evidence: Churn by plan and tenure band, last quarter.

Recommendation: Review onboarding and first-month experience
for that plan; consider a targeted plan-fit message.

Caveat: One quarter of data; correlation only.