sanash43
commented
4 months ago Feedback addressed from Milestone 1
- Files such as: .ipynb_checkpoints, .gitconfig, bashhistory, .cache, .local/share, ipython/profile_default/startup were removed (this was also noted in Milestone 2)
- Branch where all file deletions were first tested before implementing changes to main branch: https://github.com/DSCI-310-2024/DSCI-310_Group-Project_Group8_Purchasing-Intent-Analysis/activity?ref=file-deletion-test
- Implementation in main branch by @nourshawk: https://github.com/DSCI-310-2024/DSCI-310_Group-Project_Group8_Purchasing-Intent-Analysis/commit/bf634e11f738a70a90eda9dab7c8f0268bcc64f3
- All related file names are lowercase
- yml file moved to the root folder
- deletion of folder with incorrectly named venv file: https://github.com/DSCI-310-2024/DSCI-310_Group-Project_Group8_Purchasing-Intent-Analysis/commit/c8207a77c3d1b72cc2aca262bf6996aaa511ede2
- adding venv file to the root repository: https://github.com/DSCI-310-2024/DSCI-310_Group-Project_Group8_Purchasing-Intent-Analysis/commit/e67aae5552bf320f9a747aa9d8fcbb702498ee72
- Made a project title related to the project topic
_The points below all concern the written content and ordering in the analysis. If there is not a specific link for one of the bullet points below, it is covered under this commit: https://github.com/DSCI-310-2024/DSCI-310_Group-Project_Group8_Purchasing-Intent-Analysis/commit/4db24328f3e6bd27551af0502e46bfd4468e9fa9_
- Abstract revised to:
- clearly state the question being asked
- clearly report the major findings
- discuss the importance and/or the limitations of the findings
- Introduction revised to clear ambiguities in the question and explain how the methods address the different components of the question
- Dataset is described in more detail
- Methods revised to:
- read dataset directly from the source
- fixing in in the analysis file: https://github.com/DSCI-310-2024/DSCI-310_Group-Project_Group8_Purchasing-Intent-Analysis/commit/297db9d673fa8d42a343251cae3a104621912ab4
- ensuring data repo dependencies are documented:
- train/test split was moved to after the EDA so as not to break the Golden Rule
- methods extended to answer the question better: selected best performing model which takes in all the explanatory variables, then applied feature importance to the chosen model to determine which features best predict the target variable.
- read dataset directly from the source
- Results revised to:
- ensure all tables and plots have labels with proper numbering
- under each figure, there is some description to clarify the meaning of what is observed
- Discussion revised to:
- fix incomplete segments or blanks that we overlooked to fill out from the rough draft
- findings are linked back to the question
- References revised to include the missing citation of the data source
- Spelling and grammar errors were fixed
- Setup: repository name was changed to be relevant to the project name
- Couldn't find the specific commit, but the repository name was changed from "DSCI-310_Group-Project_Group8" to "DSCI-310_Group-Project_Group8_Purchasing-Intent-Analysis", and the README.md title was changed to "Predict and Classify Online Shopper Intention"
- GitHub issues are linked to milestones for Milestone 3 and Milestone 4 moving forward
Feedback addressed from Milestone 2
- Makefile revised to include all the artificats generated under the 'all' command
- Commit history of branch where all Makefile revisions were implemented: https://github.com/DSCI-310-2024/DSCI-310_Group-Project_Group8_Purchasing-Intent-Analysis/activity?ref=makefile_edit
- Reproducibility issues fixed so that the instructions in the README.md correctly guide the user to correctly recreate the computational environment
- Makefile is revised so that the analysis artifacts are in a sensibly named directory
- GitHub issues are linked to milestones for Milestone 3 and Milestone 4 moving forward
From Peer Review Feedback
- Fixed the discrepancy in the virtual environment naming in the README.md file as per @diaazavery's feedback:
- Removed PDF rendering in the .qmd file, which was faulty due to some unresolvable latex dependencies. Confirmed with the professor that just the HTML suffices. This addresses @diaazavery's feedback.
- Added comments to explain each test case for the feature importance function in
tests\test_feature_importance.pyas per point 4 in @x99i9's feedback
nourshawk
calvinyhchoi
SaiUbc
Addressing feedback from Milestone 1,2 and Peer review