25. How to Rank and Running Totals in pandas (rank, cumsum, window)

25.1 Why Rank and Running Totals

By Ravindra Bagale, trainer LinkedIn Last updated: 24 October 2026

Why. Shraddha Bagale (Pune ops) sorts amount_inr descending and pastes the sheet on WhatsApp. Amir asks: “Is BLK-10 still #1 after ties?” Ruhi asks: “What was cumulative delivered GMV by 6 pm IST?” Sort alone does not answer rank ties or running totals.

What.

Rank — place of a row among peers (1 = best if you rank high amounts first).
Running total (cumsum) — add values in order so each row shows “so far”.
Window — a moving or growing slice (last 3 days, or everything from day 1 until now).

How.

Pick the question first

Write the manager question in one line.
If it needs “#1 / #2 / tied” → rank.
If it needs “GMV so far today” → sort by time, then cumsum.
If it needs “last 3 days average” → rolling.
Always check sort order before cumsum or rolling.

Worked example A — Sort vs rank.

import pandas as pd

df = pd.DataFrame({
    "order_id": ["BLK-1", "BLK-2", "BLK-3", "AMN-4"],
    "city": ["Pune", "Pune", "Nashik", "Pune"],
    "amount_inr": [210, 210, 95, 180],
    "customer": ["Ruhi Bagale", "Amir", "Zoya", "Ravina"],
})
print(df.sort_values("amount_inr", ascending=False))

Sorting shows order. Ranking labels ties clearly (next section).

Worked example B — Running total need (fictional evening).

Raja in Nagpur sends hourly delivered GMV. The city manager wants one column: cumulative GMV from midnight IST — that is cumsum, not a new groupby.

Worked example C — Decision card.

Question	Tool
Top basket in Pune today	`rank` (within city)
GMV so far this week	`cumsum` after date sort
3-day moving average orders	`rolling(3).mean()`

Ravindra Bagale's Tip

Pehle question. Mag rank ki cumsum. Sort bhulayla naka — cumsum ulat order var galicha number deto.