About

I’m Ashrith. I built an econometric analysis of why people fly, covering markets across the Americas, Europe, Asia and Australia, all from public data. The short answer is that flying tracks income closely. My models have income in them and no ticket prices, so that answer also counts the fall in real fares that came with rising incomes. This site is the long answer.

(It began as my MSc dissertation in Artificial Intelligence at the University of Bath; it outgrew the notebooks.)

What this is

A portfolio research project on commercial aviation demand. The United States is the deep case study, built on BTS, FRED, BEA and BLS data. Beyond it, the project models demand in Australia, Canada, the EU, the UK, Singapore, Brazil, Japan, Mexico, South Korea, China and India, and ranks the markets it can on a cross-market elasticity ladder that runs from Singapore at the low end to Brazil at the high end. China is shown on that ladder but not ranked, because its test is inconclusive on a shorter pre-COVID window. Markets with no reliable estimate, India and Canada among them, are left off it.

Start with the findings, the headline results in plain English, each with a status label. The site’s labels, defined in the glossary, are confirmed, confirmed existence, provisional, inconclusive, exploratory, descriptive, and projection. Where the evidence does not pin a figure down, the finding says so.

From there, Charts is the gallery, backed by an archive of every chart.

How it’s built

The dissertation studied several decades of quarterly US domestic air travel in Jupyter notebooks and passed with distinction. The notebooks worked but did not scale, so I rebuilt them as Kedro pipelines, expanded from the United States alone to the markets listed above, and published the results as charts, each with a field note.

Everything runs on public data, through pipelines written in Python, with every threshold and lag choice specified in YAML configuration. Methods & Data has the data sources, how the work is organised, what was estimated for each question and what the study does not establish.

Where to start

Which path to take depends on where you’re coming from. New here: the findings for the headline results in plain English, then the chart wall on the home page. Technical: the elasticity ladder field note for the status of each estimate and the tests behind it, and Methods & Data for sources and limits.

Every chart on the site has a field note that says what the chart shows. Most charts that report a model’s result also carry a status label and notes on how far to trust it.

Scope

The models are tools for understanding structure and sensitivity; the site is not a forecasting service or a dashboard. The data are a fixed vintage: the home page and Methods & Data state the last quarter covered, and the site is not refreshed as agencies publish later ones. The confidence intervals reflect estimation uncertainty and carry no predictive guarantee.

Licence & use

All written content, charts, code, and derived analyses published on this site are copyright © 2024–2026 Ashrith Yerrapragada. All rights reserved. No part of this work may be reproduced, distributed, transmitted, modified, adapted, or used to create derivative works, in any form or by any means, without prior written permission of the copyright holder.

The underlying raw data are publicly available from the US Bureau of Transportation Statistics (BTS), Federal Reserve Economic Data (FRED), Bureau of Economic Analysis (BEA), Bureau of Labor Statistics (BLS), Eurostat, and the per-market national statistics and aviation agencies listed on the Methods & Data page. Those datasets are not owned by this project, and readers who want to replicate any specific analysis should obtain the data directly from the source agencies.

AirTraffickonomics is a trademark of Ashrith Yerrapragada.

Contact

For licensing inquiries or questions about the analysis: contact@airtraffickonomics.com.