Are there any examples of open data that have been scraped from "real-world" sources? To explain what I mean by "real-world", here are some (made up) examples and then some criteria:
- Train departure times from automated OCR of crowdsourced photos of train departure boards (UK readers will understand why this might be useful!)
- Train departure information from mobile GPS measurements
- Product nutrition information from OCR of crowdsourced photos of "nutrition information" on product labels (plus barcode scanning from image data)
- As above, but product prices
- Statistics about exercise habits from mobile GPS, accelerometer and temperature data
- Statistics about noise pollution from mobile microphone measurements
- Statistics about emotional state / mental health from processing photos of people on the web and analysing their facial expressions
These all involve some combination of the following criteria:
- Non-trivial processing of some sort of physical measurement source data: image processing / OCR, natural language processing, etc.
- The lack of an intrinsic need for direct user involvement in data gathering (other than consent, and app installation, if relevant). Currently with photo data, user involvement is likely required in order to take the photo and indicate that it be processed in this way, but this may change over time.
- Crowd-sourced source data gathering from mobile phone / tablet / google glass-type devices
- Source data (photos, text) fetched from the web
- Crowd-sourced data analysis (e.g. coding photographs of people by perceived emotional state based on facial expression)