Problem

Industry partner: Girteka Logistics

Girteka Logistics is the largest European asset-based transportation company, delivering more than 880 000 full truck loads (FTLs) every year, with more than 9 000 fully owned trucks and 9 800 trailers that operate in Europe. Girteka Logistics has a clear aim to be the obvious first choice for clients, colleagues, partners, the community, and shareholders.

Every day, Girteka strives to provide the best experience possible and ensure our customers' highest transportation service reliability. With more than 22 000 passionate employees, Girteka is making a meaningful impact on the future of logistics. Girteka is committed to highlighting the importance of road safety and environmental impact and raising awareness of the responsibility every one of us has towards society and our surroundings. Girteka secures profitable and sustainable growth with its partners and shareholders through state-of-the-art digital solutions that drive it towards the future.

European road freight prices and quantity dependencies: prediction models

Foreword

We want to understand if transportation prices are fair and profitable compared to the quantities agreed. Furthermore, it is important to understand the overall potential of the transportation business.

Task

To develop a decision-making model for delivery price and quantity delivered.

Suggested methods are linear regression (univariate or multivariate), panel data analysis, correlation analysis, generalized linear models, time series analysis, or any other method of data analysis, forecasting and optimization on which the team members may be expert, including bayesian and machine learning techniques.

Minimal expected results are:

1. Analysis of mutual dependencies of region/country's macroeconomic indicators (e.g., GDP, unemployment rate, inflation, stock to sales, Brent oil price, and others).
2. Analysis of dependencies between logistics company/industry indicators (average freight rate, volume of loads) on macroeconomic indicators studied in point 1 and on additional indicators that can be considered relevant by the team.
3. A multilevel and/or multivariate model for forecasting of price and of level of fairness of prices, with multiple input variables both internal to the company (measured as average freight rate, the volume of loads, distances, etc.), and some of the external factors that are not dependent on client or company (e.g., fuel market prices, GDP, etc.) studied in the previous points.

Additionally, study of the following aspects will be considered as an added value (but not mandatory):

1. A model to forecast the total demand for transportation.
2. The suggestion of some new indicators that can influence the price of transportation.

The codes (if any) used to develop the analyses and models should be provided together with the report. Coding in Python is preferred but it is not mandatory.

Data

The company will provide its internal and some external data with records of previous agreements: dates, freight rates, volumes, fuel prices, distances, routes, etc.

Data will be provided only to registered participants of the competition and must be kept confidential by the participants, by subscribing to the confidentiality agreement in the registration form.

External data not provided by the company can be freely downloaded from public web sites like Eurostat (https://ec.europa.eu/eurostat), direct download to Python or R (e.g. Quandl or yfinance), or any other sources that the participants may know.

Countries that can be considered: Spain, Portugal, Italy, France, Belgium, Germany, Holland, Luxembourg, Austria, Great Britain, Latvia, Lithuania, Poland, Slovenia, Slovakia, Croatia, and Hungary.