Ingreed
Scraping data and predict your IT salary, get rich !
DataikuAbout
The goal of this supervised machine learning project was to develop a predictive maintenance system for vehicles on Dataiku. Using sensor data, we built a model to anticipate potential failures and trigger preventive actions before actual breakdowns occur.
Dataiku allowed efficient data preparation, model training, and evaluation through a collaborative and visual interface (see flow screenshot).The project addressed two key business questions:
- 1. What are the most common factors that lead to vehicle failures ?
- 2. Which vehicles are most likely to experience a breakdown ?
Feedback
One key advantage of using Dataiku was its no-code interface, which allowed us to focus on what really matters in a machine learning project: data cleaning and model development.
The platform enabled us to perform most tasks — from data preparation to model training — through visual workflows, without having to write complex code. This significantly accelerated our progress and reduced the technical overhead typically associated with machine learning development.
At the same time, Dataiku remains highly flexible: when needed, we could seamlessly integrate custom Python scripts into the pipeline. This allowed us to fine-tune specific steps, such as feature engineering or advanced model evaluation.
Screenshots :
