Hackathon Retail 4.0

Welcome to Hackathon Retail 4.0!

We're looking for innovators that will shape the future of retail.

Galp operates a total of 1500 service stations in Portugal and Spain, most of them with a convenience store that services drivers on the road. It’s part of our purpose to continuously improve our business and provide better experiences to the customers. 

  The Goal   

With this Hackaton, we are challenging the community to propose solutions to specific problems and use cases that we think could improve our typical customer journey in the service stations. We have 2  challenges that will be further detailed in the following sections. 

To tackle these challenges, participants can expect support from Galp’s Commercial BU, Retail 4.0 team, IT&Digital and Data Office, as well as a series of benefits and prizes supported by our Open Innovation program Upcoming Energies.

💡 The Challenges

Challenge #1: Car Wash services proposal based on AI, weather rules and computer vision

Leverage Artificial Intelligence algorithms to visually identify the level of dirtiness of the customer’s cars (one or more) to grade a vehicle so Galp can suggest the customer car wash services.

Galp has a wide network of service stations in the Iberian Peninsula ( aprox. 1500), out of which several offer car washing services to visiting customers. It is our goal to increase the overall usage of this specific service, so we must find innovative and interactive ways to encourage drivers to wash their cars once they enter the stations.

We challenge the hackers community to develop an algorithm based on Machine Learning that would be able to identify if a car is dirty once it enters the forecourt of the station. This algorithm would feed off Computer Vision datasets and prediction / historic weather rules to promote carwash services to the final client. 

👉 Find out more details about this challenge on the Categories tab

Challenge #2: Optimize in-store product assortments and estimate potential sales volume

Leverage Artificial Intelligence algorithms to identify trends, handle historical data and combine various data sets of different origins to optimize in-store product assortments and estimate potential sales volume.

With such a great deal of service stations in the Iberian Peninsula , it gets hard to understand which products sell the most, at what time and in which shops, where should they be placed in-store, etc. The challenge gets even greater if we try to do successful forecasting considering other external factors such as weather, time of the year, or football matches to optimize planning and dynamically change the products positioning. Teams will need to use publicly available data (e.g.: stocks, sales, product rotation, etc) to tackle this problem.

We challenge the hackers community to come up with a solution that is capable of treating inputs from sales data, stocks, and planograms to:

• Dynamically identify what are the products that sell the most, and correlate these trends with their in-store positioning (planogram), price, and margins;
• Associate periodic sales trends with temporary events such as football matches, time of the year, weather, daytime, and other external factors that might affect sales;
• Predict future sales volumes;
• Based on these inputs, propose planograms and prices that optimize sales at each time;
• Consider the introduction of new products based on external data such as market trends.

👉 Find out more details about this challenge on the Categories tab

🥷 Audience

Anyone that is passionate about solving problems: you can be a coder, a developer, an IT market professional, or a university student.

🏆 Prizes

Winner of Challenge #1: 10.000€

Winner of Challenge #2: 10.000€

📁 Deliverables

You must register on the hackathon and start a project page by the 29^th April, only at the end of May it is necessary to submit the completed project.

By the 23rd May, you will have to publish on the platform, a detailed Project Description and the following deliverables: 

• All the assets that comprise the submission (code assets, configurations, documentation, etc) must be delivered on a branch of the github repository with a pull request submitted on branch main assigned to the team.
• At least the following information is required:
  • A brief description of the layout of the folders and their purposes / what assets they comprise
  • A brief description of the asset itself
  • A documentation (markdown, restructured text, etc) of the whole submission detailing what was the approach to solving the challenge

• The documentation delivered inside the repository should present this following topics:

  1. Delimitation about the problem

  2. Hypothesis of solution

  3. Team members

  4. Team strategy

  5. Technologic pre requisites

  6. Used data sets

  7. How the solution can be used

  8. Metrics and results

   

  1. Delimitation about the problem

  The team vision about the challenge selected and why the challenge was selected by the team

   

  2. Hypothesis  of solution

  Detailed description about how the challenge can be solved in the team perspective

   

  3. Team members

  Who and how the team organize ours roles and responsibilities

   

  4. Team strategy

  Detailed description about the plan of resolution of the challenge

   

  5 .Technologic pre-requisites

  What kind of technologic tools is needed for the solution (what galp needs to validate the functionality of de deliverable), all the tools used should be free of charges to enable Galp to evaluate, if some tools isn't available the project will be disqualified.

   

  6. Used data sets

  What data sets are used and where they can be picked. (attention to EULA and legal use about the data, ilegal data set will be disqualified)

   

  7. How the solution can be used

  A manualized application of the final product with use example.

   

  8. Metrics and results

  How the team evaluate the solution delivered and how the metrics can be afford

  Evaluation Criteria

1. The project reached the final objective of the challenge?
2. The project proves a hypothesis?
3. The experiment is reproducible (Galp can execute the code and achieve the same results)?
4. The results of the project are measurable?
5. The solution can be scalable?