### Methods Bites

Blog of the MZES Social Science Data Lab

### Using Geospatial Data in R

The use of geospatial data – data that can be mapped using geographic information systems (GIS) – has become increasingly widespread in the social sciences. Applications not only extend to the analysis of classical geographical entities (e.g., policy diffusion across spatially proximate countries) but increasingly also to analyses of micro-level data, including respondent information from georeferenced surveys or user trace data from Tweets. In this Methods Bites Tutorial, Stefan Jünger (GESIS) and Denis Cohen (MZES) show how to retrieve, manage, and visualize geospatial data in R. Continue reading

### Generalized Additive Models: Allowing for some wiggle room in your models

Generalized additive models (GAMs) have become an important tool for modeling data flexibly. These models are generalized linear models where the outcome variable depends on unknown smooth functions of some predictor variables, and where the interest focuses on inference about these smooth functions. In this Methods Bites Tutorial, Sara Stoudt (Smith College) offers a hands-on recap of her workshop “Generalized Additive Models: Allowing for some wiggle room in your models” in the MZES Social Science Data Lab in March 2021. Continue reading

### Extracting Emotions from Faces with Face++ (and Microsoft Azure)

Images are an increasingly used data source in the social sciences. One application is to extract features from human faces using machine learning algorithms. This blog post provides a guide on using APIs for this task, specifically how to access the services offered by Face++ and the Microsoft Face API. The post walks you through (1) how to gain API access credentials, (2) how to call the Face++ API from R, and (3) how to handle the output. It is based on the talk by Theresa Küntzler, who introduced the participants of the MZES Social Science Data Lab on May 12, 2020, to Extracting Emotions (and more) from Faces with Face++ and Microsoft Azure. Continue reading

### regplane3D: Plotting 3D regression predictions in R

The interpretation and presentation of empirical findings from (generalized) linear models has come a long way in the social sciences. Researchers increasingly visualize substantively meaningful quantities of interest such as expected values, first differences, and average marginal effects and consistently include uncertainty estimates in the form of analytical, simulation-based, or bootstrapped confidence intervals. However, existing interpretations and presentations are typically restricted to bivariate patterns which show (changes in) expected values as function of a single predictor, holding all else constant. This can be a significant limitation, especially when substantive inquiries focus on the interplay of two variables in predicting an outcome. To interpret and visualize such applications effectively, researchers must extend their presentations to include a third dimension. In this Methods Bites Tutorial, Denis Cohen and Nick Baumann introduce and showcase the regplane3D package, a tool for plotting 3D regression predictions in R. Continue reading

### Teaching Quantitative Social Science in Times of COVID-19: How to Generate and Distribute Individualized Exams with R and RMarkdown

The COVID-19 pandemic has forced universities around the globe to switch from on-site teaching to online teaching. As a consequence, quantitative social science classes that previously relied on closed-book in-class exams now have to administer open-book take-home exams. A downside of this switch is that it becomes impossible to monitor compliance with no-collaboration rules. Individualizing exam prompts can prevent students from sharing digital answers while taking the exam. Yet generating, distributing, and correcting individualized exams can be highly time consuming unless the procedure is automated. In this Methods Bites Tutorial, Denis Cohen, Marcel Neunhoeffer and Oliver Rittmann present an approach for the automated generation of individualized exam prompts and solution sheets, along with their automated distribution via email, using R and RMarkdown. Continue reading