MangoViz, an open-source interactive web interface to view and explore data of long-term agronomic experiments on mango
2025
Normand, Frédéric | Doisy, Anna | Grechi, Isabelle
Long-term agronomic experiments provide interesting data and enable various analyses such as yield regularity over years or effects of climate on tree productivity. However, the visual exploration of these large data sets, that precedes statistical analyses, can be a long and tricky task with several angles of approach: by variable, by factor level, by year, by specific set of years, etc. To simplify and speed up this exploration, we developed MangoViz, an interactive web interface to view and explore data of long-term agronomic experiments on mango. MangoViz is developed with the open-source 'shiny' R package and can be used from any device with internet access. It can handle data from different experiments that are processed separately. Data of a 15-year pruning experiment and a 12-year varietal evaluation are currently included in the interface. Three variables can be explored: fruit mass, number of fruit and mean individual fruit mass tree-1. For each variable, three interactive graphs are provided. The first graph shows the distribution of the selected variable for each level of the tested factor (e.g., pruning modality or cultivar) and for a selection of years. The second graph shows the changes of the selected variable, per factor level and/or per tree, across years. The third graph is the spatial distribution of the selected variable, per factor level and per year or averaged over years, on a plan of the experimental orchard. Buttons next to the graphs allow to quickly select the desired years and factor levels, and graphs are adjusted instantly. Tabs give access to useful information and documentation on the experiments. Long-term trends could be detected in the two current data sets, suggesting specific statistical analyses. MangoViz code is open-source. It can be adapted to fit users' data sets and specific questions.
显示更多 [+] 显示较少 [-]