https://github.com/carlescg/budapest-bi-2017-reproducible-bike-ride

Slides material for the talk "A reproducible bike ride in Hamburg (…or elsewhere in Europe)" at Budapest BI 2017
https://github.com/carlescg/budapest-bi-2017-reproducible-bike-ride

bike bikesharing r reproducibility

Last synced: 6 days ago
JSON representation

Slides material for the talk "A reproducible bike ride in Hamburg (…or elsewhere in Europe)" at Budapest BI 2017

• Host: GitHub
• URL: https://github.com/carlescg/budapest-bi-2017-reproducible-bike-ride
• Owner: CarlesCG
• License: mit
• Created: 2017-11-14T12:30:55.000Z (about 7 years ago)
• Default Branch: master
• Last Pushed: 2017-11-27T13:44:21.000Z (about 7 years ago)
• Last Synced: 2024-10-29T23:46:10.799Z (about 2 months ago)
• Topics: bike, bikesharing, r, reproducibility
• Size: 90.7 MB
• Stars: 0
• Watchers: 2
• Forks: 0
• Open Issues: 0
• Metadata Files:
  • Readme: README.md
  • License: LICENSE

Awesome Lists containing this project

README

        
# BudapestBI 2017 - Reproducible bike ride 

[Slides material](https://github.com/CarlesCG/budapest-bi-2017-reproducible-bike-ride/raw/master/BudapestBI2017_Reproducible_Bike_Ride_CarlesCG.pdf) for the talk *"A reproducible bike ride in Hamburg (…or elsewhere in Europe)"* at [Budapest BI 2017](http://budapestbiforum.hu/2017/en/).

## Abstract

The talk focuses on the use case analysis *“Bike sharing usage in Hamburg”*. It will discuss techniques that can take this base analysis and make it reproducible, as well as extend it to other locations in Europe. Base analysis [poster case](https://www.user2017.brussels/uploads/Kruse_poster-session.pdf).

Reproducible research is gaining traction in academia, as well as in enterprise environments. The lack of it can lead to significant issues in:  

   + Science  

   + Corporate

   + Health care

   + Policy making. 

   

Data scientists must provide not only the results, but also the data, code, and environment where the analysis is executed. After the first iteration where the application works as desired, decisions need to be made:  

   + How reproducible the report/research/analysis should be?  

   + Which data archiving strategy to use? 

   + Which data to archive? 

   + Which steps can be more cost/time effective to make the analysis reproducible in the future?

   + Follow best practices or step away from them? 

   

All those trade-offs and more are part of the craft of a data scientist. The application language is R, although same concepts are applicable to other languages, environments and use cases.

![](https://github.com/kruse-alex/bike_sharing/blob/master/bike_usage_HH.png)