BEGIN:VCALENDAR
VERSION:2.0
PRODID:-//pretalx//pretalx.com//pydata-london-2026//talk//RTWLPY
BEGIN:VTIMEZONE
TZID:GMT
BEGIN:STANDARD
DTSTART:20001029T030000
RRULE:FREQ=YEARLY;BYDAY=-1SU;BYMONTH=10
TZNAME:GMT
TZOFFSETFROM:+0100
TZOFFSETTO:+0000
END:STANDARD
BEGIN:DAYLIGHT
DTSTART:20000326T020000
RRULE:FREQ=YEARLY;BYDAY=-1SU;BYMONTH=3
TZNAME:BST
TZOFFSETFROM:+0000
TZOFFSETTO:+0100
END:DAYLIGHT
END:VTIMEZONE
BEGIN:VEVENT
UID:pretalx-pydata-london-2026-RTWLPY@pretalx.com
DTSTART;TZID=GMT:20260607T161500
DTEND;TZID=GMT:20260607T170000
DESCRIPTION:Sailors avoid the word ‘rope’. Once it has a job\, it becom
 es a line with a specific name: halyard\, sheet or warp. In forecasting\, 
 we often do the opposite — projections\, baselines\, scenarios and targe
 ts all end up being called ‘the forecast’.\n\nIn practice\, forecasts 
 live in a high-dimensional space. They vary by origin date\, prediction ho
 rizon\, scenario assumptions\, uncertainty representation\, reconciliation
  level and decision context. Treating them as a single artefact creates am
 biguity\, semantic drift and misaligned expectations.\n\nIn this talk\, I
 ’ll show how we reframed forecasting at Spotify as a structured predicti
 on problem rather than simply a modelling task. I’ll cover practical des
 ign patterns for representing forecast objects across multiple origins and
  scenarios\, handling probabilistic outputs\, implementing hierarchical re
 conciliation and tracking lineage and versioning in Python-based systems.\
 n\nAimed at data scientists and ML engineers working with production syste
 ms\, this talk offers a framework for thinking about forecast dimensionali
 ty and concrete implementation patterns you can apply in your own forecast
 ing platforms.
DTSTAMP:20260602T223122Z
LOCATION:Hardwick Hub
SUMMARY:Don’t Call It “The Forecast”: Designing Prediction Systems at
  Scale - Thomas Ogden
URL:https://pretalx.com/pydata-london-2026/talk/RTWLPY/
END:VEVENT
END:VCALENDAR