Coding and AI classes in Bremerhaven

import csv

messages = 0
ships = set()
with open("ais_positions_2025-12-01.csv", encoding="utf-8") as f:
    for row in csv.DictReader(f):
        messages += 1
        ships.add(row["ship_id"])

print("Position messages:", messages)
print("Different ships:", len(ships))
print("Messages per ship, on average:", round(messages / len(ships)))

A newcomer reading the file would say "172,306 ships". The set says 170. A set keeps each value once, however often it is added, so it turns a stream of messages into a list of distinct ships. On average each ship sent about a thousand position reports that day, because ships that move report often.

Next question: what kinds of ship were they? The type comes in the longer messages, so Version 2 builds a dictionary from ship number to type, then compares it with the set of ships from Version 1. Set subtraction shows which ships never said what they are.

from collections import Counter

kind = {}
with open("ais_ship_types_2025-12-01.csv", encoding="utf-8") as f:
    for row in csv.DictReader(f):
        kind[row["ship_id"]] = row["ship_type"]

by_messages = Counter(kind[s] for s in kind)
print("Ships that sent their type:", len(kind))
print("Seen moving or moored but never said their type:", len(ships - set(kind)))
print("Top kinds:", by_messages.most_common(4))

Thirty-four ships sent positions but no type during the day, so any chart of "ship types in Bremerhaven" built from this file is missing a fifth of the ships. Then there is that odd label, "Cargo shipA". The file also contains Cargo shipB, C and D: the letter is an extra category attached to the type. For a simple count of cargo ships they belong together, which Version 3 does.

tidy = Counter()
for s, k in kind.items():
    if k.startswith("Cargo ship"):
        k = "Cargo ship"
    tidy[k] += 1
print("Top kinds after tidying:", tidy.most_common(3))

Data: Bundesamt für Seeschifffahrt und Hydrographie, AIS-Daten zur Schiffsabgasmessung an der Messstation Bremerhaven 2025 12, day file for 1 December 2025, licence Datenlizenz Deutschland Namensnennung 2.0, "Quelle: Bundesamt für Seeschifffahrt und Hydrographie", downloaded 3 November 2026. All counts are ours. Message counts are not ship counts: one moving ship can fill thousands of rows.

1. Ages 6 to 8, no screen. Count every car that passes a window for five minutes, writing a tally mark each time. Then count how many different cars it was. Why are the numbers different?
2. From about 9, Python. Run the first version, then count how many messages said "moored" and how many different ships ever said "moored".
3. Teenagers. Find the two ships that reported two different types during the day. Which type does the dictionary keep, and how would you decide which is right?
4. Adults. Work out, for each hour of the day, how many different ships were heard. Which hour was the quietest, and does that match the message counts?