Big Data analysis and processing competition for Data Scientists
fill in the briefThe term “championship” is used to hold competitions between teams of mathematicians and data scientists most often, as a result of which the models for analyzing big data that businesses need are built.
It is logical to use the championship format for companies with a large amount of data (historical and real) to receive new models for processing and analysis. Based on the obtained models, it is possible to build forecasts, improve business processes.
These formats are often used by companies to find and attract young people and new talented data analysts to the company.
The championship format can be applied to any other competitive mechanics between industry specialists, where the results obtained by the teams are close to reality and the impact of which on the customer's business can be assessed at the current time.
Tell us about your problem, and we will figure out how to solve it!
Client: Russian Agricultural Bank
The third Data Science championship for students, which is held by Rosselkhozbank JSC and the AgroCode Hub community. Participants need to come up with a solution for finding spare parts for agricultural machinery from photographs.
← previous case
next case →
Client: VTB
Data Fusion is an international conference and competition in a concept which means combining data, merging or transferring algorithms from one area of machine learning to another, as well as the synergy of processes in machine learning. The Spinon team, together with VTB, conducted an announcement campaign for the Data Fusion and Data Fusion Contest conferences.
Over 1,000 Data Fusion Contest participants
Over 44,000 views of the Data Fusion conference broadcast
← previous case
next case →
Client: Gazprom-Media RTV
Championship among data scientists to determine the game event in sports broadcasts. The participants were required to create a method for automatically detecting game events in broadcast football matches, as well as to determine the type and time of the event relative to the start of the match (and not from the start of the video recording). Based on the solutions sent, a rating table was created. The winners who provided the models to the organizer received a prize.
Over 500 participants
← previous case
next case →
Client: PIK digital
Championship among data scientists, which was organized in two stages. At the online stage, it was necessary to develop a model for predicting sales rates for three months in advance in the context of buildings and apartments of a certain number of rooms in different projects. All solutions were included in the public rating, and the authors of the best 80 solutions were invited to PIK Digital Day, the offline stage of the championship.
850 applications
80 finalists
550 000 rubles prize fund
← previous case
next case →
Client: Sber
The aim of the Championship was to identify the bank's clients prone to churn, according to some existing features that characterize the behavior of clients. The solution of this problem allowed the bank to take preventive measures to retain customers who could potentially refuse the bank's services in the near future.
149 applications
← previous case
next case →