Does your company need help with Data projects?

The Lviv Data Science Summer School is an annual event organized by the Faculty of Applied Sciences, Ukrainian Catholic University. The school’s dates are July 20th – 31st. The venue is the UCU campus in Lviv, Ukraine. The audience includes nearly 70 participants: bachelor, master, and Ph.D. students, as well as young professionals. About one-third of the participants are from Lviv, another one-third are other Ukrainians, and one-third international students (mostly from Europe, but also from other continents).

The aim of the school is to give the participants, who already have foundations in Data Science and Machine Learning, application knowledge and skills in various domains like Computer Vision, Natural Language Processing, Game Theory, Financial Data, Reinforcement Learning, etc. You may find a whole list of the courses provided during the school on the school’s website. For those who are just starting their journey in the DS/ML field, we provide a one-week intensive introduction on Statistics, Machine Learning, and Data Visualization basics on July 13 – 18. During the second half of the school, the students work on the projects. The projects’ topics are provided by lecturers and partner companies/organizations.

We look for real-life projects from companies that are developing their own products and services, dealing with various data and analytics (or just planning to do so).

You will get a diverse team of highly-motivated students who are eager to apply their freshly obtained knowledge to your problem. On average, a team consists of three-to-four participants. As a result, you may get a validation of your data potential, prove (or deny) a working hypothesis, and get a proof-of-concept that may help to boost your project idea.

You have to provide a detailed problem statement with the following aspects described: the domain overview, the issues and the challenges you’re dealing with, the problem you want to solve, and the data available. In terms of data availability, we need to have concrete data files and/or the sources where the data could be gathered and/or expectations about the nature of the data.

Also, we expect that the company representatives will be in contact with a student team and provide their supervision during the project work period. The latter may include domain-specific guidance as well as helping with technical and methodological issues if there will be such need.

The project results are presented publicly during the last day of the school. The company representatives are welcome to visit the project presentation session and provide feedback on the students’ work.

Please beware that we do not provide a guarantee of positive results after the project work. Also, as the students select the projects they want to work on, there could be high competition for the projects and for the students. As a result, there is a non-zero chance that a project might not even get enough interest from the students and will be left without a team.

The partnership is free of charge.

Case study: Uklon, www.uklon.com.ua

Uklon is a Ukrainian company which provides an on-demand taxi service. Uklon’s analogs are Uber, Lyft, Taxify. For the Lviv Data Science Summer School 2018, the company provided a project titled “X demand forecasting. The project proposal stated:

“The aim of this task is to find optimal conditions for the fulfillment of our mission and commitments to our clients and partners. We want you to find the balance of demand and supply in our service through the most rational distribution of cars in the city at each point in time.”

The data was provided in terms of a set of taxi user requests and routes for a period of a couple of months for a Ukrainian city. The proof-of-concept solution had to forecast the demand for the taxi requests for a fixed period of time and geographical area and an individual forecast for a driver.

Case study: Ricker Lyman Robotic, rickerlyman.com

Ricker Lyman Robotic is a U.S. company with a production office based in Lviv, Ukraine. The company has a broad background in DS/ML problems from marketing to computer vision. For the school 2018, the company provided a project titled “Detecting a person’s direction of interest”. The project proposal stated:

“The main goal of the project is to detect the direction of a person’s head view from photo or video. The solution should work in real-time, so one shall start with classic computer vision methods and move to approaches based on neural networks.”

There were no labeled data provided. So in order to complete the project, the students had had to record and to label a set of video fragments.

Case study: Covizmo, http://covizmo.com/

Covizmo is a Danish company with an office in Lviv, Ukraine. The company has three partners with extensive experience within IT development, configuration of ERP systems and master data management. Covizmo offers a range of services that helps companies in their digital transformation process towards the development of digital products that support their core business. Machine Learning team has deep expertise in various areas of Artificial Intelligence. For the school 2019, the company provided a project titled “Visual inspection of wind turbine blades. The project proposal stated:

“The aim of the project is to create an instrument for detection and decrease the horizontal displacements of separate images in a stitched view of the blade side of the wind turbine.”

The data was provided like images of wind turbines parts (with annotated blade outlines) and their positions on the stitched view that recreates the original image of the wind turbines blade. The stitching of the images is performed automatically based on metadata collected by drone (position and orientation of the drone, distance from the drone to the blade at the moment when each image was made). Students proposed the solution to this problem: a cross correction algorithm.

 

Case study: EVO.companyhttps://evo.company/

EVO is a Ukrainian product IT company. Company projects are the largest marketplaces for online shopping, electronic document management, public and commercial auction. Evo helps entrepreneurs grow their businesses and make online shopping easier. For the school 2019, the company provided a project titled “ Recommendation Engine for a Procurement System (Zakupki.prom.ua).

Zakupki.prom.ua is an official participant of a public procurement system of Ukraine Prozorro and an open commercial purchase system Rialto. The project proposal stated:

The main goal of the project is to create a recommendation system to help executors find the procurements relevant for them. Target action: an executor applied on procurement from the proposed list.

The data was provided like tabular data containing information about applications on procurement by some executor behavioral information about executor information about a procurement (who posted, when, in which region, expected price, etc.). Students created a model and proposed metrics for solving this problem.

During the four previous schools, almost 60 projects were completed. The projects’ presentation slides and video recordings are available on the past schools’ web pages (at the bottom of each page):

If you’re interested in providing a project for the Lviv Data Science Summer School purposes, we would be happy to discuss the potential collaboration. Please contact us by email [email protected], Oleksii Molchanovskyi, Head of the School.