---
title: "Discussion on \"From 1.4M NYC Taxi Trips to Production ML: \nFeature Engineering Improved Predictions by 34.6%\" | Hashnode"
description: "Discussion on \"From 1.4M NYC Taxi Trips to Production ML: \nFeature Engineering Improved Predictions by 34.6%\". How thoughtful feature engineering and SHAP analysis turned 1.4 million NYC taxi trip into a production ML model with 34.6% improvement over baseline."
url: "https://hashnode.com/posts/from-1-4m-nyc-taxi-trips-to-production-ml-feature-engineering-improved-predictions-by-34-6/6a3cff0b1f1a3ea93dcea54f"
---

[Feed](https://hashnode.com/feed) /

Discussion

# [From 1.4M NYC Taxi Trips to Production ML: Feature Engineering Improved Predictions by 34.6%](https://akshat-ml.hashnode.dev/from-1-4m-nyc-taxi-trips-to-production-ml-feature-engineering-improved-predictions-by-34-6?utm_source=hashnode&utm_medium=feed)

How thoughtful feature engineering and SHAP analysis turned 1.4 million NYC taxi trip into a production ML model with 34.6% improvement over baseline.

Jun 25 · 6 min read · akshat-ml.hashnode.dev  
Read the article

-   [#machine-learning](https://hashnode.com/tag/machine-learning)
-   [#feature-engineering](https://hashnode.com/tag/feature-engineering)
-   [#lightgbm](https://hashnode.com/tag/lightgbm)
-   [#python](https://hashnode.com/tag/python)
-   [#kaggle](https://hashnode.com/tag/kaggle)
-   [#streamlit](https://hashnode.com/tag/streamlit)

## Written by

[akshat jain](https://hashnode.com/@akshat-j0in)

@akshat-j0in · Mumbai, India

  
[View profile](https://hashnode.com/@akshat-j0in)

ML projects from exploration to deployment. Building models that solve real problems.

## Responses

Markdown supported · ⌘↵ to post

No responses yet.