FANS nettverksmøte | Data Science/Analytics | 13.3.2024 | Oslo/Hybrid

[BeStibolt]

Velkommen til FANS nettverksmøte innen Data Science/Analytics.

Dato: 13/3-2024

Tid: 13:00-16:00

Sted: SAS Institute, Parkveien 55, 0256 Oslo (4. etasje). Alternativt via Microsoft Teams (linken for å delta får du i bekreftelses-eposten etter du har  registrert deg).

Det vil bli servert mat og noe godt å drikke etter nettverksmøte kl. 16.00.

Agenda: 

13.00-13.10 FANS

13.10-13.40 Hvordan utforske/preppe data for å bruke i modellering i VA

Presentasjonen holdes av Pia Rønnevik, SAS Institute.

13.40-14.10 Hvordan kan du oppdage og redusere skjevheter i SAS Viya

Presentasjonen holdes av Mathias Lanner, SAS Institute.

14.10-14.25 Pause

14.25-15.15 Data Science in search for the simplest explanation for difficulties of ski tours

Ski touring requires physical fitness, good equipment, technical skills and experience in assessing danger or difficulty in the terrain. The published difficulty level from the literature of the Swiss Alpine Club (SAC) is an important criterion for route selection. To extend this SAC metric to the entire Alpine region, the question arises if a fully automatic machine learning method can determine tour difficulties consistent with the published SAC difficulties for the Swiss Alps. Ideally such automatic method should provide full transparency about what determined the difficulty level of each ski tour (white box algorithm). We present a model trained on Swiss Data to explain the SAC difficulty levels and to adjust for the local-regional bias that is partially present. Each of the 1307 ski tours was decomposed into 10m segments, and their local topographic information such as slope, fall speed, forestation, curvature, etc. was derived from a digitized map of the landscape model.  Methods of variable selection, linear optimization, in combination with statistical techniques such as quantile regression and robust regression show that in practice, fall hazard along the ski tour are the primary factor in assigning SAC difficulty levels. Fall hazard is defined here as the local interaction between the slope gradient and the maximum speed reached in the event of a fall. The proposed model allows not only to transparently, consistently, and automatically extend the technical difficulty levels of the SAC methodology to ski touring tracks anywhere in the world, but also to localize the partial difficulties along each route on the map.

 The advantages for skitourenguru.ch are obvious:

1. consistent evaluation, i.e. reduction of a possibly existing evaluation bias.
2. interpretability of the route evaluation.
3. independence from current snow- and weather conditions.
4. efficient initial- and re-evaluation of large route datasets beyond Switzerland

For the portal Skitourenguru.ch these features are of fundamental importance for the extension of its information service to the entire Alpine region.

The presentation is held by Ulrich Reincke, SAS Institute Germany.

15.15-15.30 Pause

15.30-16.00 Oppdag det forventede og uventede ved å bruke Tekst analyse i VA

Presentasjonen holdes av Vegard Hansen, SAS Institute.

Dersom du ønsker å holde en presentasjon, så ta kontakt med Pia Skare Rønnevik.

Dette nettverksmøte er for FANS medlemmer. Hvis du ennå ikke er medlem av FANS, er du fortsatt velkommen til å delta for å se om dette er interessant for deg i fremtiden. Se FANS medlemspriser og fordeler.

PÅMELDING

Har du andre spørsmål, ta kontakt via FANS epost.

Gå ikke glipp av events og informasjon!

Husk at du unngår å gå glipp av relevante events og informasjon hvis du melder deg på det månedlige nyhetsbrevet for SAS-spesialister  (på engelsk). Du kan når som helst melde deg av.