Using Machine Learning to Solve Complex City Problem

Case Study of Sihanoukville

Case Study of Sihanoukville


Applying Machine Learning & AI for urban city development challenge

using Machine Learning and AI to provide near real-time analysis of how Sihanoukville Province change in practice.

By teaching computers to analysis different dataset gather from different source such as satellite image, drone imagery and different sensor, our machine learning algorithms can be leveraged quickly and provide pro-active information such as communities that are prone to flooding.

Screen Shot 2019-08-05 at 5.54.02 PM.png

using Machine Learning Identify & Classify Sihanoukville Waste

Mapping of waste and illegal dumping waste across Sihanoukville city and classify between Construction Waste, Household waste and Commercial Waste.

Generate report as a heat map for each different category, Identify Why and Where the waste are potential generated from.

Screen Shot 2019-08-05 at 5.55.22 PM.png

Construction Waste potential generate by construction near by that illegal dump on the empty sidewalk or infront of empty land black


Commercial and Household waste generate from new resident and commercial business setup.

using sidewalk as a waste illegal dump also block the stormwater flow contribute traffic and storm water flood


Wast generate by construction that not manage their own construction site.

Without proper wall cover their construction site and using sidewalk to put all of their construction material and waste.

This case flood due to the construction wast block the water flow.

Risk Assessment & Briefing

Located, Identify and Analysis Current Status of all Construction Activities across city

  • Located and Mapping all construction activities

  • Point out construction site issue such as: Construction Wall Cover, Construction Material located outside or on sidewalk

  • Construction wasted and environment around site

  • Construction permit and other legal requirement

  • Construction Inspection Report

Screen Shot 2019-08-05 at 10.15.23 PM.png