We performed data visualisations to interpret impact, patterns, and trends of claims. The aim of this case study was to apply visualisation techniques with the use of R and some of its main visualisation packages to present frequency and severity data for natural disasters. This was used as a proxy for loss ratios and ultimately claims in respect of these disasters, in a manner that allows the viewer to interpret the impact, patterns and potential trends of the occurrence of the disaster.
What we have done:
We streamlined the traditional approach that consisted of manual inspection, spreadsheet analysis, pivot tables, and graphs. Traditionally, the data would have to be in a set, structured format specific to the needs of each type of analysis performed, but our approach allowed for streamlined, unified data input. Our work involved:
Reworking and presenting a database containing a record of global natural disasters that occurred between 1900 to 2018.
Using a basic scatterplot, fitted in R, displaying number of disasters, deaths from disasters and economical damage from disasters recorded between 1900 and 2018 which provides an overview of the data composition and quality.
Creating a bar plot in R displaying the distribution of the number of disasters.
Creating a boxplot that illustrates the median number of disasters across all disaster types per year as well as the outliers.
Creating stacked bar plots with the same data as but with the added detail of the type of disasters. These stacked bar plots provide further details on the composition of disasters.
For this client, we have delivered:
Streamlined processes for data analysis and visualisations