Sociology student becoming a computational social scientist.
I am finishing my BA in Sociology at HSE University (2026) with a minor in UX-design.
My work sits where quantitative methods meet sociological questions. Rather than relying only on survey data, I build my own datasets from APIs and digital traces — then operationalise social concepts into measurable variables, and model them with inferential statistics. Two questions drive most of what I do: how do you turn something fuzzy like "atmosphere" or "compatibility" into a variable you can defend? and what does the measurement quietly assume?
- Residential mobility in Russia — who moves once, who moves repeatedly, and whether attitudes separate the two groups. Survey analysis with factor analysis, clustering and logistic regression, in the research group «Городская повседневность».
- Attitudes toward ghosting — a survey plus a randomised vignette experiment (48 scenarios, 1,872 ratings) on why silent disappearance is judged acceptable in one relationship and not in another.
- My BA thesis on how recreational spaces in Murmansk communicate with students during the polar night — a natural experiment in the sociology of time and urban space.
- Text-as-data pipelines: content analysis codebooks that are reproducible and auditable, not just plausible.
- Measurement critique: I document what a dataset cannot tell you as carefully as what it can.
| Project | What it is | Methods & tools |
|---|---|---|
| residential-mobility Research group «Городская повседневность» |
Who moves, and who moves repeatedly? A survey study of residential mobility (N = 7,563) — structure, attitudes and beliefs about success. | SPSS metadata · factor analysis · hierarchical clustering (city-block & Gower) · MCA · logistic regression with VIF diagnostics · Python, R, statsmodels, factor_analyzer, FactoMineR |
| ghosting-attitudes HSE University, 2025 |
Why is silent disappearance judged acceptable in one relationship and not in another? Survey (N = 468) + randomised vignette experiment (48 scenarios, 1,872 ratings). | Factorial survey design · cumulative-logit GLMM with random intercepts · Wilson CIs · Spearman ρ · Mann–Whitney U · Kruskal–Wallis H · VIF diagnostics · SPSS, Python |
| polar_night BA thesis, HSE, 2026 |
Text-as-data study of ≈9,230 VK posts from Murmansk venues: does the language of urban space shift when the sun stops rising? | VK API · 44-variable codebook · χ² tests · 6 logistic regressions with OR, 95% CI, McFadden R², AUC-ROC · Python, statsmodels, Jupyter |
| murmansk Data infrastructure |
Automated hourly panel of real-time venue occupancy: 11,696 observations across 355 venues, collected by a GitHub Actions cron job. | Google Places API · spatial sampling frame · panel-data design · GitHub Actions · Python |
| studyBuddy UX-design minor, 2026 |
Full-stack app matching students into study pairs, with a weighted multi-criteria compatibility model and LLM-based semantic similarity. Built by vibe coding with an AI agent. | Next.js · tRPC · MySQL · Drizzle · Jaccard & overlap similarity · Llama 3.1 via Groq · Vitest |
Quantitative: survey and questionnaire design · factorial vignette experiments · survey data analysis · multilevel / mixed-effects models (GLMM) · computational text analysis · content analysis and codebook design · factor analysis · hierarchical clustering · multiple correspondence analysis · logistic and ordinal regression · χ² tests of independence · nonparametric tests (Mann–Whitney, Kruskal–Wallis, Spearman) · collinearity and VIF diagnostics · model evaluation (OR, 95% CI, McFadden R², AUC-ROC) · similarity metrics (Jaccard, overlap, Gower, ordinal distance) · panel-data construction
Code: Python (pandas · NumPy · SciPy · statsmodels · scikit-learn · factor_analyzer) · R (tidyverse · gtsummary · FactoMineR · gower · marginaleffects) · SPSS · SQL / MySQL · TypeScript · React / Next.js · Git & GitHub Actions · REST API data collection · LLM APIs
Languages: Russian (native) · English (academic)
- Email: m.balabina52@gmail.com
- GitHub: @mbalabina
Open to research assistant positions, thesis collaborations and anything at the intersection of quantitative sociology and data science.