Written by @jamiewildehk
Currently this is only scraping the thunderstorm warnings, though it could easily be extended to hot/cold weather warnings and typhoon warnings.
Only one end point now: /hko/ts/[date:YYYYMMDD]
The scraping is performed by Yahoo Pipes, using the YQL defined in the YQL Query I wrote
Implemented as an express application for heroku or own hosting. Similar code could be used client side.
{
"warning": {
"warnings": 1,
"totalDuration": 60,
"details": [
{
"from": "2014-07-14T21:00:00.000Z",
"to": "2014-07-14T22:00:00.000Z"
}
]
}
}