Obtaining, Analysis and Visualization of Mobility and Traffic data in Quito – Ecuador

The increase of traffic in large cities cause serious problems, such as traffic accidents, delays at workplaces, stress and other problems, it makes it necessary to use new technologies to identify these zones and make a better decision on vehicular circulation. This work proposes to collect data using a Drone, analyze this data and predict traffic in a sector from patterns, using traffic and weather photographs. Using Python code, through an ETL the data is extracted from a non-relational database, then loaded into MongoDB, the downloaded images will be passed to an object detection tool that will be used to locate vehicles in each image, a type of traffic (heavy, medium, low, empty) is assigned based on the number of vehicles found in the image, with this data, predictive models can be tested. The model will finally perform traffic forecasts using data visualization tools to better reflect the days and times of traffic flow in the sector. The initial results show great effectiveness in identifying areas with high vehicular flow, which will allow an adequate management when circulating in these areas.