Data Analytics for Supply Chain & Operations India 2026
Demand Forecasting, Inventory, Logistics & Your Career Transition

# Simple demand forecast in Python with Indian festival seasonality flag

import pandas as pd
import numpy as np

# Load weekly sales data
df = pd.read_csv('weekly_sales.csv', parse_dates=['week_start'])
df = df.sort_values(['sku_id', 'week_start'])

# 13-week moving average forecast
df['ma13'] = df.groupby('sku_id')['units_sold'].transform(
    lambda x: x.rolling(window=13, min_periods=4).mean()
)

# Flag Indian festival weeks (Diwali, Navratri, New Year)
festival_weeks = ['2026-10-12', '2026-10-19', '2026-10-26',  # Navratri/Dussehra
                  '2026-11-02', '2026-11-09',                 # Diwali
                  '2026-12-28', '2027-01-04']                  # New Year
df['is_festival_week'] = df['week_start'].isin(pd.to_datetime(festival_weeks))

# Forecast error (MAPE)
df['forecast_error'] = abs(df['units_sold'] - df['ma13']) / df['units_sold']

# Flag high-error SKUs needing attention
high_error = df[df['forecast_error'] > 0.20].groupby('sku_id')['forecast_error'].mean()
print(f"SKUs with >20% forecast error: {len(high_error)}")

-- Inventory ABC analysis + aging in SQL
-- Works on ERP exports from SAP, Oracle NetSuite, Tally

WITH sku_revenue AS (
    SELECT
        sku_id,
        sku_name,
        category,
        SUM(revenue)                                    AS total_revenue,
        SUM(revenue) / SUM(SUM(revenue)) OVER ()        AS revenue_share,
        SUM(SUM(revenue)) OVER (ORDER BY SUM(revenue) DESC) /
            SUM(SUM(revenue)) OVER ()                   AS running_pct
    FROM sales
    WHERE sale_date >= CURRENT_DATE - INTERVAL 90 DAY
    GROUP BY sku_id, sku_name, category
),
abc AS (
    SELECT *,
        CASE
            WHEN running_pct <= 0.80 THEN 'A'
            WHEN running_pct <= 0.95 THEN 'B'
            ELSE 'C'
        END AS abc_class
    FROM sku_revenue
),
inventory AS (
    SELECT
        sku_id,
        SUM(stock_qty)                                  AS total_stock,
        MAX(last_movement_date)                         AS last_move,
        DATEDIFF(CURRENT_DATE, MAX(last_movement_date)) AS days_no_movement
    FROM warehouse_stock
    GROUP BY sku_id
)
SELECT
    a.sku_name, a.category, a.abc_class,
    ROUND(a.revenue_share * 100, 1)                     AS revenue_pct,
    i.total_stock,
    i.days_no_movement,
    CASE WHEN i.days_no_movement > 90 THEN 'SLOW MOVER' ELSE 'OK' END AS stock_status
FROM abc a
LEFT JOIN inventory i USING (sku_id)
ORDER BY a.abc_class, a.total_revenue DESC;

-- OTIF (On Time In Full) analysis by carrier and region

SELECT
    carrier_name,
    delivery_region,
    DAYNAME(promised_date)                                  AS day_of_week,
    COUNT(*)                                                AS total_shipments,
    SUM(CASE WHEN delivered_on_time = 1 AND qty_delivered = qty_ordered THEN 1 ELSE 0 END)
                                                            AS otif_count,
    ROUND(
        100.0 * SUM(CASE WHEN delivered_on_time = 1 AND qty_delivered = qty_ordered THEN 1 ELSE 0 END)
        / COUNT(*), 1
    )                                                       AS otif_pct,
    ROUND(AVG(DATEDIFF(actual_delivery, pickup_date)), 1)  AS avg_transit_days,
    ROUND(AVG(delivery_cost_inr / distance_km), 2)         AS cost_per_km
FROM shipments
WHERE promised_date >= '2026-04-01'   -- Indian FY start
GROUP BY carrier_name, delivery_region, DAYNAME(promised_date)
HAVING total_shipments > 50           -- filter out low-volume noise
ORDER BY otif_pct ASC;                -- worst performers first

-- Vendor performance scorecard in SQL

SELECT
    v.vendor_name,
    v.vendor_category,
    COUNT(po.po_id)                                         AS total_pos,
    ROUND(AVG(DATEDIFF(gr.receipt_date, po.promised_date)), 1)
                                                            AS avg_delivery_delay_days,
    ROUND(100.0 * SUM(CASE WHEN gr.receipt_date <= po.promised_date THEN 1 ELSE 0 END)
          / COUNT(*), 1)                                    AS on_time_pct,
    ROUND(100.0 * SUM(gr.rejected_qty) / NULLIF(SUM(gr.received_qty), 0), 2)
                                                            AS rejection_rate_pct,
    ROUND(SUM(po.po_value_inr) / 1e5, 1)                  AS total_spend_lakhs,
    ROUND(100.0 * SUM(po.po_value_inr)
          / SUM(SUM(po.po_value_inr)) OVER (), 1)          AS spend_share_pct
FROM purchase_orders po
JOIN goods_receipt gr ON po.po_id = gr.po_id
JOIN vendors v ON po.vendor_id = v.vendor_id
WHERE po.po_date >= '2026-04-01'
GROUP BY v.vendor_name, v.vendor_category
ORDER BY total_spend_lakhs DESC;

Is supply chain analytics a good career in India in 2026?

Yes — supply chain analytics is one of the fastest-growing specialisations for data analysts in India in 2026. The growth is driven by three forces: the rapid expansion of Indian e-commerce (Flipkart, Meesho, Amazon India) which requires real-time inventory and logistics analytics; the Indian government push for manufacturing under PLI schemes bringing international supply chains to India; and post-COVID supply chain disruptions accelerating digitalisation in traditional sectors like FMCG, pharma, auto, and consumer goods. Supply chain analysts with both domain knowledge (understanding what OTIF, fill rate, and lead time mean) and data skills (SQL, Excel, Power BI) command a premium over pure technical profiles at the same experience level.

What data analytics skills are needed for a supply chain analyst role in India?

For a supply chain analyst role in India, the core analytics skills are: SQL (for querying inventory, order, and vendor databases — the most universally required), Excel (SUMIFS, pivot tables, XLOOKUP — still the daily workhorse in manufacturing and FMCG companies), Power BI or Tableau (for building logistics dashboards and KPI reports shared with management), and Python with pandas (for demand forecasting models, large dataset processing, and automation). Domain knowledge of SCM metrics — OTIF (On Time In Full), Days of Inventory Outstanding (DIO), Fill Rate, Order Cycle Time, COGS — is equally important and is your advantage over a pure data analyst who has no SCM background.

Can a supply chain or operations professional switch to data analytics in India?

Yes — and this is one of the most natural transitions in the Indian job market. Supply chain professionals have three advantages over typical data analyst freshers: they already understand the business context behind the data (what a stockout means, why lead time variance matters, what a 3PL does), they have exposure to real operational data and the messy reality of ERP systems, and they understand what analytics output is actually useful to decision-makers. The gap to bridge is technical — SQL, Python or Power BI, and the ability to structure an analysis. A supply chain professional who learns these tools and applies them to their own domain knowledge typically reaches interview-ready in 4–5 months and commands a 15–25% salary premium over a generic data analyst with the same experience.

What is demand forecasting analytics and how is it used in India?

Demand forecasting analytics is the process of predicting future product demand using historical sales data, seasonality patterns, and external factors — so companies can plan inventory, production, and procurement accordingly. In the Indian context, demand forecasting must account for Indian seasonal patterns (festival demand spikes during Navratri, Dussehra, Diwali, and New Year; monsoon impact on FMCG sales; wedding season effects on categories like appliances and apparel). Methods range from simple moving averages (Excel) to exponential smoothing (Python with statsmodels) to machine learning-based forecasting (using sklearn or fbprophet). Even a basic moving-average forecast that is maintained and monitored consistently is far more valuable than no forecast, which is the current state in many Indian SME supply chains.