Automating report generation to improve the work of World Shops

The Challenge

The World Shop Umbrella Association and the One World Network Bavaria (hereinafter referred to as the World Shop Umbrella Association) conduct annual surveys of their members—individual World Shops—to better understand their situation and needs. Additionally, members should receive beneficial feedback regarding their position compared to other World Shops.

The survey aims to provide each participating World Shop with individualized feedback. The corresponding report contains various graphics and descriptive statistics, enabling individual World Shops to assess their situation relative to others. Before the project, these individualized reports were manually created using Excel and Word. With several hundred participating members, this was a highly time-consuming and error-prone process, which the predominantly volunteer staff of the World Shop Umbrella Association struggled to manage.

The World Shop Umbrella Association sought to automate this process. Furthermore, statistical analyses were to uncover additional insights into the structure of World Shops.

The Data Basis

The data were collected through an online survey conducted among World Shops. The dataset included several hundred participating shops and dozens of variables.

The data were exported from the online survey tool as .csv files and provided to the CorrelAid team. During the project, the data were stored exclusively on the Umbrella Association’s systems and CorrelAid servers, ensuring no unauthorized third-party access.

The Approach

CorrelAid supported the World Shop Umbrella Association by automating the manual, error-prone, and time-intensive creation of individualized reports. The open-source statistical software R and the LaTeX typesetting system (implicitly via R) were used to generate tables within the reports.

As part of the project, proposals for optimizing the questionnaires were also developed and implemented.

To ensure long-term impact, the Umbrella Association’s team was empowered to conduct the analysis independently in the future. The scripts developed by CorrelAid were discussed in detail and tested extensively with the Umbrella Association’s staff. The code was also thoroughly commented. The project team remained available for follow-up questions, which the Umbrella Association utilized repeatedly. The process now runs smoothly on the staff’s local machines.

Technical Solution Approach

The data were imported and processed in R (data quality control, aggregation, creation of new variables). Descriptive analyses were performed for all shops as well as for individual shops and their respective peer groups. The results were used to generate the individualized reports, which were created using Rmarkdown. Tables in the reports were produced with the R package *kableExtra*, and graphics were created using *ggplot2*. The entire process can now be triggered for all participating World Shops and runs automatically.

The Impact

Thanks to the automation approach, the World Shop Umbrella Association can now create reports more easily, sustainably, and with significantly less effort. Additionally, data validity has improved.

Through the project, the Umbrella Association gained a better understanding of the general approach to data-driven evaluations. This includes aspects such as designing surveys with future data analysis in mind and considerations for data quality assessment. The tools used and their capabilities were also conveyed.

In turn, the CorrelAid volunteers gained a deeper understanding of the needs and workflows of NGOs in general and World Shops in particular. Moreover, they learned extensively from one another during the project—both in terms of tools and collaboration methods.