Conclusion

When looking at all of these models and results, many technical conclusions can be drawn regarding the best model for the data, the forecasting accuracy of the model, the important features in the model, and the results from submissions to the competition. However, the primary goal was to better understand the data collected from a survey intended to identify beneficiaries eligible for government assistance for housing reconstruction.

It was important to identify the most critical factors when it came to building damage. When rebuilding, these factors would be taken into consideration so that the damage from future earthquakes would be greatly reduced. Given this, conclusions regarding factors affecting household damage were garnered, such as:

1. The structural materials used for buildings matters. Structures engineered using reinforced concrete were far safer than structures built out of other materials. Conversely, structures built using flagstone were far more susceptible to damage from the earthquake.

2. The location of the building versus the damage sustained needs to be acknowledged. However, since it is virtually impossible to predict the location or size of the next earthquake, this variable is not very effective.

3. When assessing buildings in the future, it is important to note the age, height, and area of the building. Perhaps buildings of certain age, height, or area could be further reinforced for future protection.

A building inspector should be able to expand the use of this project to gain insight into a building's susceptibility to earthquake damage. Along the same lines, a building inspector could use this information to create stricter building codes based on height and areas of buildings that sustained more damage than other buildings.

Using this project as a starting point, this could go in many directions. First, using exploratory data analysis, a further in-depth study could reveal more critical relationships between sets of variables and their impact on the damage of buildings. From these discoveries, sets of building codes could be constructed to maximize the safety of buildings in preparation for the next earthquake. Considering earthquakes are not unusual in that region, the next earthquake is bound to cause damage to buildings that were similarly damaged in the previous earthquake.