Week ten was an experimental week for me here at Othot. We are testing new ideas for the customer data I have been working with since the beginning. After the meeting, they communicated some specifics they would like to see us do with their data. So this week, we took a look at them. I had to merge some new files together which was a lot more easy than anticipated. After merging, I cleaned it up a little bit by removing unnecessary columns that would not help us make predictions. After engineering it, I created a preprocessing script, similar to the one we used with the old data, and ran it too make sure it worked. Everything panned out okay, then we moved onto making predictions. A brand new algorithm created by one of my superiors, makes this process more efficient. Before, our output files would all be jammed into one single file, with the new algorithm, we can now look into each separate file making it easier to find what we want!
I ran the scripts over and over again and taking notes on what we think looks the best. The results are decent, and I think with more data they would be more sufficient.
I give myself a quality performance rating of 5 this week because I was able to merge and write a preprocessing script on my own. Also, I am able to interpret the results effectively from what I have learned here at Othot.