An adaptive agent for Google Place crawling

EasyChair Preprint 934, version 2

Abstract

Intelligent agents are used in different academic and professional areas for various scope, one of this is Marketing and Social Media. With this work is described an adaptive agent for Google Places crawling based on Earth space honeycomb  tessellation developed to resolve the first question of my Master Thesis Master: crawling all Google Place of an urban area in order to feeds geospatial marketing analysis.

In the context of this work the urban area is the real environment while the Google Place API is a digital representation of it where the agent goal is capturing all places of an area with a minimum input.

The agent, in completely autonomy and with a minimum input, capture all places starting from a central point, by means of a spiral movement up to a maximum diameter, both specified by user. This spiral movement (spiral-pattern) it is made over an honeycomb tessellation.

The agent behaviours are characterized by: planning of movement path, collecting and storage of places, checking of results, if necessary adapting of it behaviours, fault tolerance and replanning of actions. All of this by the minimum user-input are composed by: center of crawling, default size of cells and finally the number of spirals of crawling.

The core of algorithm is the adaptation on some environmental details of his planned track and granularity of tessellation previously planned. The algorithm choose when use more smaller cells and where, and, if there are some problem, where re-planning cells. 

Keyphrases: adaptive agent, geospatial, social media, urban analysis

@booklet{EasyChair:934,
  author    = {Domenico Monaco},
  title     = {An adaptive agent for Google Place crawling},
  howpublished = {EasyChair Preprint 934},
  year      = {EasyChair, 2019}}