Team:Heidelberg/Tempaltes/iGEM42-W-10b

From 2013.igem.org

Scraping of BioBrick count

There is an API provided for accessing the registry, but unfortunately it only allows for search of a single specified BioBrick id. The fastest, but still very slow way to connect a team with it's BioBricks is doing an API request for every id in the teams parts range. The only assumption we made in order to speed things up, is that the BioBricks were submitted in a continuous parts range. Thus the matching is aborted when the first BioBrick in the parts rang can't be found.

Implementation of Scoring

In order to solve the very sensitive problem of rating a team's success we started out with a subjective scoring by every one of our team members for the different awards and the medals. This native scoring turned out to be pretty consistent an thus we just calculated mean values and put the on an exponential scale in order to achieve a harsh separation of the highest top scoring teams and those who didn't get that far. For every team the score of the single awards was added up. As the awards rewarded differend every year, we normalised the summarized score of the teams in one year to a scale from 0 to 100%.
The whole analysis was added to the R-script converting the JSON file to the RData file.

Further text analysis and data conversion

In order to do the text analysis for the methods extraction we collected a raw list of methods, which is displayed in table 10.1. They were clustered in Preprocessing, Processing and Analysis. The script doing the analysis in python again stems both the abstract and the methods and matches them.

The RData file was extended and now also includes the meshterms within the data-list as well as the information content and the track in the data-frame.