liveBook · Manning

Movement data

As explained in Investing for Programmers

Movement data records the locations, trajectories, or activity patterns of people, vehicles, or other objects over time.

Movement data uses geolocation, satellite imagery, or webcam imagery to reveal what is happening in an area over time. In algorithmic trading, it can provide underrepresented signals about companies, products, or deployments. section 11.1.1

A practical movement-data workflow can include:

  • Collect imagery or location information from relevant sources. section 11.1.1
  • Apply computer vision to count objects, such as vehicle brands on streets. section 11.1.1
  • Track changes, such as the growth rate of Waymo vehicles in particular areas, and interpret them as possible evidence of business success. section 11.1.1

Waymo deployment signal

Counting Waymo cars across locations could help estimate deployment growth. The signal may be informative, but collecting imagery and training models to identify vehicles correctly can require substantially more effort than analyzing mainstream data. section 11.1.1

How Investing for Programmers covers it

How deep it goes
Defines the signal and sketches collection, computer-vision analysis, and interpretation, but stops short of implementation details.
Examples
Counts vehicle brands in street imagery and tracks Waymo vehicle growth across areas.
What it assumes
Assumes access to imagery or geolocation data and computer-vision skills for reliable object identification.
What you can do after
Evaluate company activity through geographic signals while weighing data-collection and model-training costs.
How it connects
Connects movement data to geolocation, satellite imagery, webcam imagery, computer vision, and alternative data.

This explainer was generated by Manning's knowledge graph from the text of Investing for Programmers. The section links point to the parts of the book it draws on.