Taxi Trips – 2022 dataset from the City of Chicago open data portal#

Stage 2 · 🔗 Associations & Correlation · Lesson 08 of 56 · beginner

◀ Previous · Discovering Associations Through Data: From Everyday Patterns to Chicago Taxi Trips (September 2022) · Next · Objective Selection of the Bin Width for a Time Histogram ▶ · ↑ Section

Important

✨ AI-generated content. This page was written with the assistance of an AI language model and is provided as a learning aid. Despite careful review, it may still contain mistakes, omissions, or out-of-date information. Whether you are new to the topic, a team lead, or a senior practitioner, treat it as a starting point rather than an authoritative reference: read it critically and independently verify anything you act on (code, commands, figures, and factual claims) against official documentation and primary sources before relying on it.

An open, real dataset#

The running dataset is the Taxi Trips file from the City of Chicago open data portal — a real, public record of taxi journeys the city collects as a regulator. It is a favourite teaching set because it is large, genuinely messy, and full of intuitive relationships to explore.

What’s in a row#

Each row is one trip, with a unique ID and a timestamp, described by around two dozen fields. The most useful for association work are the numeric ones — Trip Miles (distance), Trip Seconds (duration), Fare, Tips, Tolls and Trip Total — alongside categorical fields like Payment Type and Company, and location fields (pickup / dropoff community area, census tract, and centroid latitude / longitude). Most columns load as floats or text.

Scale and loading#

The full record runs to hundreds of millions of trips over the years; a single week of September 2022 is already a workable subset of tens of thousands. It is published through the city’s Socrata portal, so you can download a CSV and read it with pandas.read_csv (or pull a filtered slice via the Socrata API) rather than loading everything at once.

It needs cleaning#

Being real, it needs the preparation this course is about. Records appear with trip end before start, absent durations, impossible distances (over 100 miles), fares below the city’s base charge, and missing community areas for trips outside Chicago. Filtering these out is a prerequisite before any association or model is trustworthy — a concrete instance of why data preparation dominates the workflow.

Hint