Sample Page

Data Viz
Published

September 10, 2025

One-liner: Interactive ridership analysis with time-of-day patterns, station flows, and anomaly spotting.

Stack: Python, Polars, Altair, Quarto

Overview

Brief summary of the problem, your approach, and results. Keep it tight—bullets work great:

  • ETL from CSV → Parquet; fast transforms with Polars
  • Feature engineering for weekday vs weekend patterns
  • Altair charts for linked brushing across time & stations

Hero Image / Plot

Overview dashboard

Highlights

  • 35% speed-up vs prior Pandas pipeline
  • Interactive flows map shows rush-hour asymmetries
  • Clean Quarto layout, deployable with AWS Amplify

Code Snippet (Python)

import polars as pl df = pl.read_parquet(“data/metro.parquet”) daily = df.group_by(“date”).agg(pl.col(“rides”).sum()) daily.head()