Peer Review

[micahkwok]

Hi Group 28,

I was assigned to your group for peer review. Just some small thoughts below.

Make File

• I tried starting from scratch by using make clean and then make all, but there are some packages that I didn't have such as chromedriver_binary. I don't think this is a problem but more of a heads-up for your docker step. I didn't install all the packages and just proceeded the review by restoring the files so if there are any issues in the make file after that I can't say for sure.
• Documented well as I also saw that one of the scripts had a comment that it takes 2.5 hrs to run, so I didn't run the scripts to test
• The make clean function worked perfectly

hotel_cancellation_predict_report.Rmd

• Very minor but I thought it may be better to list all team members as authors for your report
• It's mentioned that there is class imbalance, but I thought it would be nice to know the proportions for it as I was quite curious. I know this information is available in the EDA report but I was coming from the approach of someone just reading your final product.
• Tables 2 & 3 could have more formal column header names (eg Classifier Name instead of classifier_name)
• Overall, it was quite easy to follow along and I understood the analysis, good work!

[debanandasarkar]

Hi @micahkwok ,

Thanks for taking time and looking into our project. With respect to Make file, I believe once we have our Dockerfile, things will be easier. However if you want to run it before Milestone 4, I would suggest you use conda environment of 573 and install altair_saver and chromedriver_binary on top of that. If you are using 573 environment, you may get some error in the last step as it uses R to render our final report. It depends on whether your R is set up to run from 573 or not. To quickly overcome this, you can switch back to base environment and run make all again to run the last part.

Cheers, Debananda