Accurate rebuild cost assessment (RCA) is crucial for insurance providers, customers, and government agencies involved in setting relevant building standards. RCA has relied heavily on the expertise and judgment of surveyors, which can be costly and may lead to inconsistent results, particularly in the case of less experienced surveyors new to the industry. The emergence of machine learning (ML) methods offers a promising solution to address these challenges. This paper is the first to explore the use of industry-scale data combined with explainable ML approaches to advance the field of ML-based RCA. We compiled a dataset of over 83,000 properties surveyed by experts and developed XGBoost and Random Forest models trained on this dataset. Extensive experiments were conducted to evaluate key performance metrics, including root mean square error, mean absolute error, coefficient of determination, and a business-focused measure, the Percentage Mean Absolute Error Ratio, along with ablation studies based on property types and surveyor expertise. The comprehensive results demonstrate that the performance of the developed ML models is comparable to that of human surveyors. As potentially the first study of its kind using realworld survey data for development and validation, this research provides unique insights for the insurance industry, property owners, and government agencies to enable more reliable and efficient insurance cost estimation and decision-making.

From 2018 to 2023, Brazil experienced its most fiercely contested elections in history, resulting in the election of far-right candidate Jair Bolsonaro followed by the left-wing, Lula da Silva. This period was marked by a murder attempt, a coup attempt, the pandemic, and a plethora of conspiracy theories and controversies. This paper analyses 437 million tweets originating from 13 million accounts associated with Brazilian politics during these two presidential election cycles. We focus on accounts' behavioural patterns. We noted a quasi-monotonic escalation in bot engagement, marked by notable surges both during COVID-19 and in the aftermath of the 2022 election. The data revealed a strong correlation between bot engagement and the number of replies during a single day (r=0.66, p<0.01). Furthermore, we identified a range of suspicious activities, including an unusually high number of accounts being created on the same day, with some days witnessing over 20,000 new accounts and super-prolific accounts generating close to 100,000 tweets. Lastly, we uncovered a sprawling network of accounts sharing Twitter handles, with a select few managing to utilise more than 100 distinct handles. This work can be instrumental in dismantling coordinated campaigns and offer valuable insights for the enhancement of bot detection algorithms.