Żabka shares its knowledge! This time we are inviting students to use Big Data in practice. We are waiting for projects whose conclusions will bring real financial benefits to our company or social gains for the town.
Build a three-strong team of students. Develop your project concept together and submit the entry on our website. The competition jury will select 5 best projects to be implemented with the use of publicly available data or data provided by Żabka.
1 - 31 March 2019
As one of five selected teams, implement your project, for example in R or Python using Big Data analysis tools (e.g. Hadoop, Spark), with the support from Żabka supervisors.
5 April - 16 June 2019
Enjoy the main prize together with your friends: a trip to Silicon Valley plus pocket money for the trip, or cash prizes awarded to 2nd - 5th place winners.
Results announcement: 27 June 2019
Required components of your competition project
IT - description of the analytical environment configuration in the cloud
- create the analytical database (analytical cluster), pull data for analysis
- run data transformation
- optimise the database structure for speed of query execution
Data science - the way of analysing the socio-economic problem using AI methods
- apply statistical tools (e.g. R, Python) for data analysis
- build the analytical model and analyse the economic and/or community problem
- explain your way to present the problem solution (e.g. PowerBI, Shiny)
- explain the benefits for the Organiser / community
Examples of research problems which your project can solve
- Product group or shop group price sensitivity analysis – based on historical sales data (incl. price, sales volume, promotional campaigns) you need to build a model to compute price flexibility for a product, product group or shop.
- Sales substitution analysis between product categories / products – you need to build a model to show what happens to the sale of Tyskie beer during a promo campaign for the Żubr brand, how it will affect the whole beer category and sales generated by an entire shop.
- Sales forecasting for shops – you need to build a model to forecast total sales (PLN) for a shop in the following month, taking into account temperature, shop historical sales, environment.
- Sales forecasting for promotional products – a model to forecast promotional product sales based on previous promo campaigns (the sale of similar products). The model should take into account product price flexibility and marketing flexibility (TV adverts, posters).
- Analysis of customers using the Żabka application – you need to build a model for app customers management – customer segmentation, analysis of promotional activity effectiveness by segment, promotional campaign personalization.
- Shop notifications of delivery time – build a model to compute an estimated time of arriving at a shop (one vehicle delivers to several shops). The model should take into account data on road repairs, temperature forecasts, precipitation, traffic jams, unloading time in the previous shop.
- Automation of shop orders – develop a model to generate a shop order based on the shop’s current stock of goods, historical sales, as well as weather and seasonal factors, and promotional campaigns.
- Community problems (own data), e.g.:
• Locational - optimum distribution of municipal bicycle stations / car sharing vehicles / kindergartens / schools,
• Road conditions in Poznań – optimum network of cycle lanes to minimize the amount of traffic lights and to ensure the most consistent routes for such lanes (the least necessity to cycle on the road),
• Medical (any).