Política Brasileira entry #4
Today I've serialised the script, cleaned the notebook and re-written some functions for readability. Right now I'm not sure what other information I should keep, so I'm uploading the file to Tableau to start visualising it. Due to some missing records on the original file(the first 50 entries do not have Name of the Congressman or Party) I've downloaded a file with all the congressman that has ever taken a chair. That might sound excessive, but it is good information to have. I'll have a look at the file tomorrow. All of the following functions can be applied on Tableau, so I'm gonna serialise the script to pass on the previous years (4 years) to better look at things. I still haven't figured out which Machine Learning Model I'm gonna deploy, but I had an idea today that might settle it: the question I'm trying to answer is who would I possibly vote(I'm not currently voting). So, if I deploy an unsupervised model, to cluster different profiles of Congressmen(using, here, just how much they've spent on the past four years, analysing movement trends) that could be a start. In order to do that, I'll first visualise it on Tableau then try to concatenate the remaining data frames. I'm adding another eight pomodori to the tally because yesterday I didn't count them.