How Pandas Helps You Land a Business Analyst Job Faster

import pandas as pd

# Load a sales export (CSV or Excel)
sales = pd.read_csv("sales_export.csv", parse_dates=["order_date"])
# sales = pd.read_excel("sales_export.xlsx")  # needs openpyxl installed

# First look, like scrolling the top of a sheet
print(sales.head())
print(sales.shape)          # (rows, columns)
print(sales.dtypes)         # column types
print(sales.isna().sum())   # blanks per column
print(sales.describe())     # min, max, mean for numbers

# Excel habit                ->  pandas equivalent
# Filter: Region = North      ->  sales[sales["region"] == "North"]
# Sort by Amount, largest     ->  sales.sort_values("amount", ascending=False)
# SUMIF by region             ->  sales.groupby("region")["amount"].sum()
# COUNTIF status = Returned   ->  (sales["status"] == "Returned").sum()
# VLOOKUP customer name       ->  sales.merge(customers, on="customer_id", how="left")
# Pivot table                 ->  sales.pivot_table(index="region", columns="month",
#                                     values="amount", aggfunc="sum")
# Remove duplicates           ->  sales.drop_duplicates()

import pandas as pd

sales = pd.read_csv("sales_export.csv")

# Standardise text: ' north ', 'NORTH' -> 'North'
sales["region"] = sales["region"].str.strip().str.title()

# Convert types safely; bad values become NaN/NaT for review
sales["order_date"] = pd.to_datetime(sales["order_date"], errors="coerce")
sales["amount"] = pd.to_numeric(sales["amount"], errors="coerce")

# Look at problem rows before deciding what to do
bad_rows = sales[sales["order_date"].isna() | sales["amount"].isna()]
print(len(bad_rows), "rows need review")

# Remove exact duplicates and cancelled orders
sales = sales.drop_duplicates()
sales = sales[sales["status"] != "Cancelled"]

# Revenue by region and month, plus month-on-month growth
sales["month"] = sales["order_date"].dt.to_period("M")

monthly = (sales.groupby(["region", "month"])["amount"]
                .sum()
                .reset_index())

monthly["growth_pct"] = (monthly.groupby("region")["amount"]
                                .pct_change() * 100).round(1)

# Pivot for a manager-friendly view
report = monthly.pivot_table(index="region", columns="month",
                             values="amount", aggfunc="sum")
print(report)

# Save for stakeholders (requires openpyxl)
report.to_excel("revenue_by_region.xlsx")

import pandas as pd

orders = pd.read_csv("orders.csv", parse_dates=["order_date"])

# Last order date and total spend per customer
cust = (orders.groupby("customer_id")
              .agg(last_order=("order_date", "max"),
                   total_spend=("amount", "sum"),
                   orders=("order_id", "nunique"))
              .reset_index())

# Customers with no order in the 90 days before the latest date in the data
cutoff = orders["order_date"].max() - pd.Timedelta(days=90)
inactive = cust[cust["last_order"] < cutoff]

# Focus on valuable customers first
top_inactive = inactive.sort_values("total_spend", ascending=False).head(20)
print(top_inactive)

import pandas as pd

events = pd.read_csv("funnel_events.csv")   # columns: user_id, stage

stages = ["visited", "signed_up", "added_to_cart", "paid"]

# Unique users reaching each stage, in funnel order
funnel = (events.groupby("stage")["user_id"]
                .nunique()
                .reindex(stages))

funnel_df = funnel.to_frame("users")
funnel_df["conversion_from_prev_pct"] = (
    funnel_df["users"] / funnel_df["users"].shift(1) * 100).round(1)

print(funnel_df)
# The lowest conversion_from_prev_pct marks the weakest step

ANALYSIS PLAN (write before touching the data)

Business question : ________________________________
Who is asking     : ______________  Decision they need to make: __________
Data sources      : [ ] orders.csv  [ ] customers.xlsx  [ ] other: ______
Time period       : ________ to ________
Definitions       : Active customer = ____________
                    Revenue = ____________ (before/after discounts?)
Cleaning rules    : Remove cancelled orders? Y/N   Test accounts? Y/N
Outputs           : [ ] Summary table  [ ] One chart  [ ] Recommendation
Deadline          : ________

FINDINGS NOTE (one page for stakeholders)

Question   : Why did Q3 revenue fall in the North region?
Data used  : Orders Jan-Sep, cancelled orders excluded

Key findings
 1. North revenue fell compared with Q2, unlike other regions
 2. Most of the fall came from the Accessories category
 3. Repeat orders from existing customers dropped noticeably

Recommendation
  Review Accessories stock and pricing in North; run a
  win-back campaign for inactive repeat customers.

Limitations : Returns data not included; two weeks of
              August exports were incomplete.

RESUME BULLETS (adapt to your own real work)

- Cleaned and merged sales and customer exports using pandas,
  documenting every rule applied to remove invalid records
- Built a monthly revenue-by-region report in pandas that
  replaced a manual Excel process with a single re-runnable script
- Identified inactive high-value customers using order-date
  analysis and proposed a targeted win-back campaign
- Analysed a four-stage sign-up funnel and recommended
  simplifying the step with the lowest conversion

Rule: only list what you actually did and can explain.