ELCA is in possession of an exclusive, unique dataset on the Swiss used car market, comprising around 150,000 data points of cars in various conditions, types, and price ranges, advertised in Swiss retail websites.
This dataset gives us an unparalleled opportunity to analyse the market, discover trends, and eventually develop a precise price estimation system capable of easily and conveniently giving price estimates to used car sellers, something absent from the Swiss market (to the best of our knowledge). Each data point contains a car’s make, model, year, mileage, transmission and fuel type, plus unstructured text explaining the ad.
The goal of this project is to first clean, prepare, and explore our in-house dataset in order to clarify and visualise its characteristics and various trends, and develop a price prediction/estimation algorithm, using the state-of-the-art machine learning techniques such as Gradient Boosting and shallow neural networks. You will have to analyse different features, assess their relevance, and choose the most informative subset of them. Due to the abundance of nominal features, you may use vector-embedding techniques, which adds new opportunities for visualisation and inference, as well as the prediction task.
Challenges: Building a non-trivial price estimator is very difficult “in the real world” due to various reasons, from insignificant features, to un-measurable/unavailable factors. Most pricing processes and time series follow the Martingale property, which roughly means it is extremely difficult to perform better than an almost-random estimator. Finally, the cardinality of the data and the string cluster property makes it difficult to apply traditional regression methods.
Project applications: Marketing, advertisement, Retail industry
What you will learn: You will be a junior data scientist, developing your skills in machine learning, experiencing in the full pipeline of a data science project
Keywords: (fast) Deep learning, Gradient Boosting, neural networks, linear regression, transfer learning, online learnin
In this role
In this project, the goal is to:
Develop used car price estimation system for the Swiss market, plus visualisation of the market data.
What we offer
Join our team as intern and you will find a young, dynamic and culturally diverse working environment.
Your initiative is a chance ! We're constantly looking for talented individuals. Check our latest job opportunities !