Co-location Pattern Mining with Event-Centric Model using PostgreSQL in Jakarta

Data is available in vast quantities nowadays, including spatial data, which has practical, necessary, and implicit knowledge. We need mining spatial data to obtain valuable and important information using a particular method. Data mining provides methods for mining extensive data, and we have Spatial Data Mining for spatial data. Co-location pattern mining is one of the essential methods in Spatial Data Mining, where there are some approaches. One of the approaches is the Event-Centric Model. We use Jakarta Province, Indonesia data. We also use PostgreSQL with PostGIS extension to save, process and analyse the data. Besides finding the Co-location pattern rule size 3, this paper compares ST_DWithin and ST_Distance, two methods in PostgreSQL with PostGIS, to determine which is faster in implementing the Co-location rule size 3. We find ST_DWithin is faster than ST_Distance, with a margin of $\mathbf{7 9. 8 0 \%}$. We also find that in Co-location rule size 3 in Jakarta province, Indonesia with 10 major spatial features and 120 Co-location, Mosque - School - Hospital is the highest participation index, with 0.89, which means that 89% of Mosques are in the neighbourhood with School and Hospital.