Search

Background Data on people experiencing homelessness often come from time- and labour-intensive cross-sectional counts and surveys from selected samples. This study uses comprehensive administrative health data from emergency department (ED) visits to enumerate people experiencing homelessness and characterize demographic and geographic trends in the province of Ontario, Canada, from 2010 to 2017. Data and methods People experiencing homelessness were identified by their postal code, designated as "XX." Outcomes included the number of people experiencing homelessness stratified by year and week, gender and age plotted annually, the location of each ED visit, and composition changes in demographics and geographic distribution. Results Over seven years, 39,408 individuals were identified as experiencing homelessness. The number of ED visits increased over the study period in all of Ontario. The average peak in the number of visits occurred annually in September, with the fewest visits in January. Rises in overall homelessness were secondary to increases in working-age homelessness. ED presentations were concentrated in urban centres. The total proportion of patients experiencing homelessness became less concentrated in Toronto, decreasing from 60% to 40% over the study period, with a shift toward EDs outside the city. Interpretation This study shows that administrative health data can provide comprehensive information on demographics and other characteristics analyzed over time. Surveillance can be conducted cost-effectively, and changes can be tracked in real time to allow for services to be coordinated and implemented in a timesensitive manner. [ABSTRACT FROM AUTHOR]

[Display omitted] • Mineral prospectivty mapping for Au. • Random forest algorithm for integrating geoscience data. • Knowledge-based algorithms. • Superior tectonic province. This paper uses the random forest (RF) algorithm to produce two maps of orogenic gold prospectivity (MPM) across the entire Superior geologic province of Ontario and Quebec, Canada. One MPM of this entire study area is based on a compilation of mapped faults while the other is based on a manual interpretation of lineaments from airborne magnetic data. In addition, we have created four MPMs over the Abitibi region as the MPM map of the entire Superior is characterized by a mapping bias due to more intense mapping of faults over the Abitibi portion of the study area with respect to the rest of the Superior province. There are three RF maps generated over the Abitibi region based on faults, fault density and fault intersection density, faults + magnetic data and faults + gravity + magnetic data. To reduce the effect of the fault mapping bias, we developed a fourth map based on a knowledge driven (weighted sum technique). Statistically the best MPM is based on the faults, gravity and magnetic data. The major predictors of gold are NW-SE, NE-SW, EW trending faults, fault intersection density between EW and NW trending faults and to a lesser extent fault density and a magnetic vertical gradient image. In addition, we compare two greenstone belts in three dimensions with respect to their potential for gold exploration. We conclude that the Larder Lake greenstone belt is more fertile with respect to gold mineralization than the Swayze greenstone belt due to deeper penetrating faults which have greater potential for tapping gold mineralized fluid from the mantle or lower crust. Our final MPM map has identified prospective areas that contain known gold mines as well as areas that do not contain any known gold mines. These prospective areas may be prime for orogenic gold exploration. [ABSTRACT FROM AUTHOR]