Level 6 · Our top cities, counted properly

One city, many spellings

customers["city"] was typed in by people, and it shows. New York appears as New York, NEW YORK and new york, sometimes with spaces around it. To pandas those are different values, so any count splits the city into pieces:

customers["city"].nunique()                              # 129
customers["city"].str.strip().str.title().nunique()      # 61

.str.strip() removes spaces at both ends, and .str.title() capitalizes the first letter of each word and lowercases the rest. Together they map every spelling onto one clean name. Checking nunique() before and after is a quick way to see how much the cleaning merged. Look at the cleaned values once, too: title() can mangle names like “McAllen” into “Mcallen”. In this table the 61 cleaned names, St. Louis included, all come out right.

Count, then rank with a tie-breaker

After cleaning, a groupby count per city gives the numbers Daniel wants:

city	raw count	cleaned count
New York	175	181
Los Angeles	123	124
Chicago	102	106
Dallas	78	83

The raw column undercounts every big city, because the variants are counted separately. The cleaned ranking also changes places further down: Philadelphia (64) moves ahead of Miami (63).

Phoenix and Boston both have 58 customers, right at the edge of the top ten. Sorting by n_customers alone would leave their order to chance. Adding city as a second key, ascending, makes Boston ninth and Phoenix tenth every time, and Atlanta (57) just misses the list.

Missing cities

14 customers have no city at all. They show up as NaN, and groupby leaves NaN keys out by default, which is what Daniel needs: you can’t deliver same-day to an unknown city. If you ever need to count them, customers["city"].isna().sum() gives 14.

Common mistakes

• Counting before cleaning, so New York shows up as several smaller cities.
• Using .str.upper() or .str.lower() and then mismatching the expected names: the ticket wants title case.
• Calling .strip() without .str, which fails on a Series.
• Forgetting the tie-breaker on city.