PARKING

Identification of patterns and prediction of parking in the city of Lisbon to improve surveillance efficiency.

 

Objective
To create new models either to predict or to generate viable alternatives for illegal parking in the city.


Research question

What are the characteristics of the locations where exists an higher number of occurrences regarding irregular parking?

Where and at what time of day is expected to exist an higher number of illegalities regarding irregular parking?

 

Challenge

 

Challenge Brainstorming Session

 

Data understanding

Information about occurrences of irregular parking for 2018 and 2019 in Lisbon registed by the municipal police (60 000 occurrences).

 
Image
Image
Image
Image

 

 

 

Less abusive parking in Easter, Summer and Christmas vacations season.

Image
Image

 

 

Abusive parking during weekdays between 8h to 11h. Less concentrated between 14h and 19h.

 

 

Abusive parking during weekdays between 8h to 11h. Less concentrated between 14h and 19h.

Image

 

Abusive parking in Campo Santa Clara (56 occurrences).

Image
Image

Data preparation: Machine Learning

 

 

Irregular parking occurrences along with contextual data were aggregated at census tract level.

Image

Data preparation: Time Series

 

 

 

Irregular parking occurrences data was aggregated at street level in time bins of 3 hours.

Image

Main characteristics of the locations where exists irregular parking

Correlation between contextual data within each census tract and irregular parking.

Capturar2

 

Modeling: Machine Learning

Image