Comment
Author: Admin | 2025-04-27
Execute. The experiments were conducted using three different configurations. The configurations and their corresponding execution times are shown in Table 1.Table 1 Hadoop implementation results Full size table In local mode, the job is executed on the local machine without contacting the Hadoop ResourceManager. In cluster mode, the job is submitted to the ResourceManager which is then responsible for scheduling it to any compute nodes available. It is clear from the resuls that Hadoop incurred a significant overhead when a job was executed on the cluster via the ResourceManager.Spark implementationFor this implementation a similar setup was used. The configurations used are shown in Table 2.Table 2 Spark implementation results Full size table Compared to Hadoop, Spark incurred significantly less overhead when submitting jobs on the cluster. This is likely due to inexpensive data access operations. Therefore, Spark proved to be more appropriate for the selectd Bristol open dataset.Application resultsThis section presents the results of the analyses from an application point-of-view and illustrates how they can be beneficial for urban planners. The indices obtained from the experiments are plotted in Figure 6.Figure 6 Citizen perception regarding crime and safety vs economy and employment. Full size image The figure shows the variation in the responses from citizens over the years for the two indicators. The results show that between 2005 and 2012, public approval for economic and employment opportunities declined with an upward turn in 2012. On the other hand public approval of the crime and safety situation improved in 2007 but dipped back down in subsequent years. Since we have such data available for individual wards, we can calculate such trends for each ward as well. It should be noted that the questionnaires evolved over the years so their length and type of questions varied from year to year. Therefore, the overall index is just the average value of the positive responses for every question that was available.The Pearson’s correlation calculated between both indicators based on the available data was 0.2305 [30]. This is the most common measure of statistical correlation between two datasets. The value of this measure is always between 1 and -1. 1 indicates a strong correlation, 0 indicates no correlation, and -1 indicates a strong negative correlation. The value 0.2305 indicates that there is a weak correlation between the two indicators in question. The positive value indicates that improvement in the economic and employment opportunities improves the
Add Comment