JaredSplinter
commented
3 years ago Hi Rachel,
Thank you for your feedback! I added an extra example of what I meant by human behaviour in the final report and tried to explain some more on why we chose the hyperparameters we did. We also made the tables more human readable!
Hi! Hope you are all well. Here is my peer feedback:
Documentation: I think the project is very well documented, I can see straight away in the READme your final results.
Code: The code in the different script files is a little difficult to read because there aren't many comments, it could be helpful to add in little comment titles to what each code chunk does.
Analysis and reasoning: I think it would be nice to explain why the specific hyperparameters were chosen for optimization (why n_estimators and min_sample_split). Also, I do not quite understand how human behaviour could be a limitation in the conclusion of your project's results. I think it makes sense if you are describing the methods by which the data was collected, but human behaviour in my opinion shouldn't affect the actual data analysis portion unless there is some type of bias in how the models or hyperparameters were chosen (but I think these were chosen based on scores?). Maybe adding a bit more explanation would be good!
Communication: I think the report was easy and clear to follow. It would be nice to have one final concluding sentence or section about what these results mean in relation to the project question asked though.
Suggestion: To make some of the tables more human readable in your final reports, you could change the variable names (ex// table 2 change fit_time to Fit Time).
Overall, looks awesome! :)