NLP-based thesis supervisor allocation using TF-IDF, cosine similarity, and Genetic Algorithm, considering lecturer capacity, student preferences, and research topic alignment.

Panel data regression of the Food Security Index (IKP) across 22 districts/cities in East Nusa Tenggara, using life expectancy, GRDP, and food production as predictors. Covers Chow, Hausman, and Lagrange Multiplier tests to choose between pooled, fixed effect, and random effect models, followed by classical assumption checks. Code and results behind the article published in JSTAR (Jurnal Statistika Terapan), Vol 6(1), 2026.

Bayesian ARIMA forecasting of daily shallot prices in Papua (Dec 2024 - May 2025). Covers interpolation of missing days, stationarity tests, model selection, MCMC estimation, residual diagnostics, and test-set accuracy against a naive benchmark.

Spatial regression analysis of the poverty rate across 38 Indonesian provinces, using household access to electricity, safe drinking water, sanitation, adequate housing, and unmet health service need as predictors. Covers choropleth maps, comparison of distance-based spatial weight matrices, Moran’s I, Lagrange Multiplier tests, and SAR/SEM/SARMA/SDM models selected by AIC. Pair project with Salsa Rifda Erira.

Predicting house prices from listing data (bedrooms, bathrooms, land and building area, garage, location). Covers data cleaning, log transformation, variable selection with Lasso, and model comparison using train/test RMSE.

Spatial regression analysis of the percentage of poor population across 33 districts/cities in North Sumatra, using education, per capita expenditure, and life expectancy as predictors. Covers choropleth maps, spatial weight matrix comparison (contiguity and distance-based), Moran’s I, Lagrange Multiplier tests, SAR/SEM/SARMA models, and direct/indirect effects.

Exploratory factor analysis in R on sales performance and test score variables. Covers standardization, correlation heatmap, scree plot, factor extraction (minres), communalities, varimax rotation, factor scores, and model fit with RMS overall. Group assignment.

Time series forecasting of daily TransJakarta passenger counts (Jan 2023 - Jun 2024) using an Autoregressive Distributed Lag (ARDL) model with route adjustment as an external variable. Covers optimal lag selection (AIC), diagnostic tests, train/test evaluation with MAPE, and multistep forecasting.

One-way ANOVA for a completely randomized design in R, comparing chlorophyll content in cassava, spinach, and lettuce leaves. Builds the linear model in matrix form (X’X, X’y, reparameterization), computes the ANOVA table manually and checks it against aov(), then tests normality and homogeneity of variance.

Analysis of a two-factor experiment in a randomized block design using R. Tests the main effects of variety and fertilizer, their interaction, and the block effect with ANOVA, and visualizes the interaction with an interaction plot.

Multiple linear regression on provincial data (34 provinces) to identify factors affecting the democracy index, such as press freedom, number of CSOs, HDI, and public information openness. Covers multicollinearity check (VIF), outlier, leverage and influence diagnostics, classical assumption tests, and model selection (forward, backward, stepwise, best subsets).

Exploratory data analysis of earthquake depth in R. Compares the standard boxplot with the adjusted boxplot (medcouple-based) for skewed data, and checks the distribution with a histogram, density curve, and Shapiro-Wilk test.

Practice in data manipulation with R’s tidyverse, using the built-in quakes dataset (1,000 earthquakes near Fiji). Covers filter, select, group_by with summarize, arrange, and mutate to subset, summarize, and create new variables.

Introductory practice in R data wrangling using the built-in iris dataset (150 flowers, 3 species). Covers glimpse, filter, select, group by with summarize, arrange, and mutate, plus chaining steps with the pipe operator.