This project aims at predicting a building's damage during an earthquake. Using data from Kathmandu Living Labs and the Central Bureau of Statistics of Nepal, this group tries to find useful regression models to predict a building's damage after an earthquake. At this stage, the group performed initial data cleaning with some insightful exploratory data analysis and trained a single ordinary regression model.
Things I liked:
I like how your use of the heat map to indicate the correlation between different variables. This gives readers a concrete sense of the data that you are using.
I like how you generated many useful histograms and other descriptive diagrams to back up your data cleaning process and give insights into your data. Your explorative data analysis is amazing!
I like how you did many pre-processing steps such as normalization and standardization to ensure that your model will work smoothly. This is very important for future work.
Areas of Improvement:
There was not much information about hyperparameter tuning, which is an important but painstaking process in these data analyses projects.
There might be bias in the dataset about Nepal. Have you considered generalizing your model to buildings in other countries as well?
What do you expect the Bayes Optimal Error to be for your problem? It would be better if you could compare your results to the Bayes Optimal case and see how well you did.
This project aims at predicting a building's damage during an earthquake. Using data from Kathmandu Living Labs and the Central Bureau of Statistics of Nepal, this group tries to find useful regression models to predict a building's damage after an earthquake. At this stage, the group performed initial data cleaning with some insightful exploratory data analysis and trained a single ordinary regression model.
Things I liked:
Areas of Improvement:
Overall great job!