International Summer School on CSS Methods (HSE-CSS 2020)

Social Network Analysis (SNA)

Social Network Analysis is a powerful and widely used methodology in Social Sciences. This approach uses network metaphor for studying structure and features of social, economic or any other kind of relations between actors. With the emergence of massive online data, social network analysis became even more important, being widely applied for Computational Social Science research. The course covers both theoretical and methodological foundations of SNA, and practical part. We will discuss types of research questions one can answer via SNA, examples of research designs from different disciplines, and will get experience of empirical network analysis with R, learning to process network data, different ways to visualize it (including interactive), main SNA metrics and methods (centralities, community detection, density, etc). We will also discuss and practice hypotheses testing on networks and conclude with the exploration of more complex statistical models. The course will focus on acquiring conceptual understanding of SNA rather than advanced statistical modelling, and we will widely use visualization and exploratory SNA.

Text Mining

The goal of the course is to provide students with a toolkit to analyze large amount of texts using quantitative evidence. Using R, we will learn to collect, processes and analyze text as data. We will discuss such topics as: text as data, vectorization, preprocessing, stemming, lemmatization, text classification, sentiment analysis, topic modeling, Word2vec, Doc2vec.

Machine Learning in Social Sciences

Machine learning is an approach that allows building models from data, in other words, when you do not have fully pre-specified rules for making decisions and predictions but have sufficient data. These methods are now widely applied in many areas, including social science, where they used to uncover patterns embedded in data and improve model predictions. The course discusses the logic of machine learning, how it differs from the traditional analytical approaches and what they have in common. You will learn what «overfitting» means and why you should care about it, what is the difference between inference and prediction, why there is no «one-fits-all» method and how to find the best one for your problem. We will discuss the general ideas and principles of machine learning on the introductory level and will not go deeply into the math behind it. This course offers the basic toolkit to enable you to understand, evaluate, and build your own machine learning models for research and applied problems. All examples will be given in R, so you should have at least minimal working knowledge of this software. 

Bad Data

For reliable results of analysis, the quality of data is a key feature. However, quite often sources of data in computational social science are not perfect. In this course, we will discuss the main issues with typical CSS data, including web data scraping and cleaning, extraction and normalization, dealing with missing, and contradictory data, combining big and small data sources, including expert estimates.
We will discuss the main data issues and matching strategies, data diagnostics and cleaning, remedy strategies for typical data problems, including strategies for missing data: multiple imputation, FIML, EM, tree-based approaches. etc. We will as well cover the topics of unbalanced and biased samples, and strategies of working with different types of imperfect data: index construction, discretization, weighting, transformation to and methods for weaker data scales. We will work with the real-life (messy) data and together prepare it for the consequent analysis. Some useful R packages developed to describe the data, analyze its structure and prepare it for further analysis will be demonstrated.

Advanced Data Visualization

The aim of the course “Advanced Data Visualization” is to teach the attendants to create informative, easy-to-read, and eye-catching visualizations. Attendants will learn how humans’ visual perception works and how to implement this knowledge into the design of visualizations. On the course, attendants will acquire skills of advanced data visualization with the focus on R and ggplot2 and will get a general overview of visualization tools that can be applied to statistical, textual- and network data. We will cover creating publication-ready figures, deep customization of figures to highlight important details and facilitate decision making, various graphical representations of relationships involving nominal, ordinal, numeric data, probability and uncertainty, as well as approaches to visualize various kinds of unstructured data. In the second part of the course, we will focus on graph composition, both in publications and in interactive dashboards, reflecting on principles of goal-oriented communication with multiple plots, tables and text elements. We also discuss how the target audience influences the designer’s choices, and how cognitive science results can help support decision making. At the end of the course, the attendants will be granted with the opportunity to apply acquired skills in the visualization project and will create the poster or dashboard.

Machine Learning: Causal and Interpretable Models

As machine learning models are increasingly being employed to make decisions in different settings, it is important to ensure that researchers and users correctly understand and trust these models. In many cases, machine learning models tend to produce more accurate predictions that’s why they are widely used. But their complex structure makes them harder to interpret, even impossible to explore factors that influence predictions. The course is focused on state-of-the-art methods of interpretable machine learning, their limitations and opened questions. We will discuss connections between interpretability and causality, and fairness. You will learn how to explore the model, find the biases in data and provide actionable data-based decisions. The course will also emphasize on various applications which can benefit from model interpretability. Participants are expected to be familiar with basic machine learning and programming tools (all examples will be given in R).
 

The deadline for applications is April 5, 2020. Notifications about decisions will be sent to participants no later than April 10, 2020.

Invited participants will be asked to register by April 15, the latest, providing important information about the necessity of visa invitations etc.

Students of the Methods Track will need to pay the non-refundable registration (application processing fee) of RUB 1,000, we well as the participation fee of RUB 14,000 (approx. EUR 200) for one module. Students and faculty members of the Russian educational organizations (including HSE) are eligible for the reduced fees of RUB 700 and RUB 10,000 accordingly.

Students of the Research Track will need to pay the participation fee only, which is RUB 3,500. Students and faculty members of the Russian educational organizations (including HSE) are eligible for the reduced fee of RUB 2,500.

All fees include tuition, coffee-breaks during the School and printed materials. Lunches, as well as visa, travel and accommodation expenses are covered by participants.

Fees can be paid by a credit card (the link to the secure payment system is sent to the invited participants).

Participation in the Methods Track courses is open to all students (BA, MA, PhD), as well as senior researchers and professionals. However, for successful and productive participation participants need to meet the courses prerequisites. The basic prerequisite is the basic skills in R software usage. Student may take MOOC courses or other options available to prepare for the courses.

Participants of the Methods Track may apply using this application form, indicating the courses they prefer to take during the 1st and / or the 2nd module. Students are asked to provide the 2nd preference for the courses, in case the best preferred course is overbooked or cancelled.

Participants of the Research Track are invited based on the evaluation of their research proposals. Short papers (6-8 pages in the Springer LNCS format) are to be submitted using the EasyChair system. Each submission is reviewed by the organizers of the School (at least three reviews per submission).

Submissions must have the following elements: (1) Introduction – motivation, research question and problem statement; (2) Literature review of related studies; (3) Theory and proposed hypotheses of the research; (4) Research Design – proposed data and methods; (5) Preliminary findings (if applicable); (6) Conclusion – future steps of the research.

Most foreign citizens need a visa to come to Russia. To participate in the School, you must have a valid study (ordinary student) visa. Higher School of Economics issues visa invitations for those students who have registered and paid all necessary participation fees. No extra charges are taken for visa invitations. Visa guidelines for HSE international visitors can be found here.

Please mind the following essential rules before the application for visa invitation:

Your travel document (passport) must be valid for at least 1,5 years from the planned expiry date of the visa you intend to obtain (in case of HSE – CSS 2020: 21 July 2020).

We may issue a visa invitation for the dates of the School only, i.e. from July 9 to July 21, 2020.

If you live in a country other than your country of origin, please note you might be not allowed to apply for the visa in the country of your residence, or additional application requirements may apply (e.g. you might be asked to provide notary translation of your residence permit in the country of application). Please check with the Russian Embassy in the country of application.

Processing of the visa invitations takes up to 7 working days for the citizens of the EU (except the UK and Ireland), and up to 35 working days for other countries. To ensure timely issue of the invitation, we ask to send the necessary documents by April 15 the latest.

You will need to apply for visa in the country stated in the visa invitation. Please, plan your visa application process well in advance. Visa invitations to participants will be sent from June 1 to June 15, provided the applications are received by April 15.

Due to the federal regulations, all foreign citizens must register their visas upon arrival. We strongly do not recommend our international guests with study visas to take accommodation in private apartments due to many complications in visa registration. A better option is to book a hotel or reside in the HSE dormitory, where registration is done automatically.

Students may apply for accommodation at one of the HSE dormitories. Since the number of vacant places in the dormitories is limited, requests for accommodation are considered on the "first-come, first-serve" basis.

There are plenty of hotels and hostels near the School venue, offering accommodation at various prices. Hostels with affordable prices can be also found in the city center on the Nevsky avenue, which is only 2 metro stations from the HSE campus. At the same time, as July is a high season in St. Petersburg, we recommend participants to book accommodation well in advance!