Datathon at Data Science Week

Now Big Data and Data Science in general are penetrating more and more companies, and many interesting business problems are expanding that can be solved with the help of data and their processing. You can join this and try on data analysis hakatons, which have recently become increasingly frequent, many have heard about Microsoft machine learning hakaton , some participated in Deephack , in the open source data from MLClass.

Something similar will happen on the dataton, which will be held at the event called Data Science Week from August 29 to 30.

Official partners of the dataton: HeadHunter , Ozon.ru and 3data .
')
The first two provide data for analysis and formulate tasks. A company 3data provides all the necessary infrastructure for convenient work on datatone.

Now about the most interesting: about tasks.

• Prediction salary for jobs
  Probably, it is difficult to find a person who would never use the HeadHunter service and who would not face the fact that not all vacancies have a salary. The ability to predict salary from the job description would allow the applicant to show a job with an unspecified salary, but probably falling under his salary expectations.
  
• Related Searches
  Unfortunately, not all people speak the same language. And this is not about Russian, English, Chinese. And about the fact that employers can call vacancies in one way, and the applicant can look for the same vacancies, but according to some other words, formulations. In order to help the applicant to find the vacancies he is interested in, but not found by the first query he has compiled, it is necessary to solve the problem of determining similar search queries.
  
• Recommendations of rare goods. Distribution tails
  Very easy to recommend a product that is already popular. The conversion of such a recommendation will be high, but it will be useless from a business point of view. In literature, this is called a banana trap. It is much more interesting to recommend something from rarely purchased goods. This will be the task.
  

You will need to come to dataton with your laptops, from which you can go to the 3data-deployed cluster with the pre-installed Spark and Jupiter Notebook, as well as all the necessary python packages.

In general, it's great that instead of a standard technical solution - to deploy the infrastructure in some foreign cloud, the organizers attracted a Russian company as a partner. Of course, at 28 o'clock it was easy to get along with the cloud, but if we talk about production solutions, even for startups, the clouds are becoming less and less attractive due to the course, and because of the personal data law, they are also uncomfortable.

See you on datatone. Register for the event here .