Use it or Lose it: Facilitating the Use of Interactive Data Apps (IDAs) in Psychological Research Data Sharing

To demonstrate the use and versatility of IDAs, we capitalize on two open-source Python frameworks, namely, Dash (https://dash.plotly.com/) and Gradio (https://gradio.app/). Both frameworks enable users to present and share their research data in a highly interactive and easily understandable manner. Once implemented in Python, the IDA can either be hosted locally by using the Terminal on macOS/Linux systems or the Windows Command Prompt or externally, for example, on www.pythonanywhere.com or using Hugging Face Spaces. Whereas the former is especially useful during development, the latter allows for easily sharing the IDA with others, for example alongside research papers.

The first application (http://franziskausee.pythonanywhere.com/; dash demo.py) is a data dashboard that allows users to easily explore and visualize data of a psychological behavioral experiment. Within the IDA, the data is presented in a table-like format with columns referring to variables/measures of interest and rows to combinations of participants and experimental conditions (see Figure 1). The interactive features of the IDA comprise data filtering, data sorting, and inspection of specific rows/columns. In addition, user-defined figures can be created by selecting specific variables and the type of figure from different dropdown menus (see Figure 2, Figure 3). Moreover, additional information on selected variables is displayed below each dropdown menu, facilitating the understandability of variable names. The second application (https://huggingface.co/spaces/FranziskaU/gradio demo; gradio demo.py) is an implementation of a simplified automatic text preprocessing function (see Figure 4, Figure 5) as used in natural language processing (NLP) research. Specifically, user-defined text inputs are preprocessed such that sentence-specific information, such as the number of punctuation symbols, is provided, thus enabling researchers to interactively test the functionality and trustworthiness of source code published alongside NLP research papers.

The experimental dataset encompasses behavioral data of a novel approach-avoidance paradigm. For further information, please refer to https://osf.io/bsg5c/.

Open Data and the use of Open Source Software are two key principles of Open Science. However, the mere online availability of data and source code does not guarantee their reuse by other researchers (Kim & Yoon, 2017). At the same time, sharing large data sets in an understandable and transparent format that motivates researchers to explore said data sets remains a fundamental challenge (Ellis & Merdian, 2015). Interactive data apps (IDAs) have the potential of making scientific data sets more accessible and attractive, both within and beyond the academic research community. Specifically, in times of information overload, soaring time constraints, and often underdeveloped programming skills, IDAs may increase researchers’ willingness and capability to engage with large data sets and reuse them efficiently.