Understanding Earthquake Damage Prediction in Turkey

Earthquakes can be really scary. They can shake everything up and cause a lot of Damage. In Turkey, earthquakes happen all the time because of the way the land is shaped. In fact, thousands of earthquakes hit Turkey every year! In 2023, one earthquake tragically took the lives of more than 61,000 people. This is why it’s super important to figure out how much damage earthquakes can cause before they happen.

#Why Predicting Damage is Important

When an earthquake occurs, knowing how much damage it might cause can help everyone prepare. This includes planning where to send help and how to keep people safe. We looked at things like how strong the earthquake is and how well buildings are built. By putting all this information together, we can have a better idea of what to expect.

#What We Did

To understand earthquake damage better, we gathered a bunch of information. We looked at past earthquakes, how powerful they were, and how deep they happened. We also checked out the condition of buildings and how many people live in the affected areas. We used computer programs to help us predict how many people might get hurt or worse during future earthquakes.

#The Data

We collected data from earthquakes before 1950, which may sound old but was necessary for our study. This data included details like how strong the earthquake was, how deep it struck, and how many people were affected. We also looked at population density, which means how many people live in a certain area. If a lot of people live where an earthquake hits, the damage can be more serious.

#Our Approach to Predictions

Instead of using the usual method of assessing errors, we decided to use a method called Mean Absolute Percent Error (MAPE). This helps us see how far off our predictions are in a way that makes more sense. We also used another measurement called Mean Absolute Error (MAE), which allows us to understand our predictions better.

#Choosing Our Models

We tested several computer models, each using different types of data. Think of it like trying on different outfits at a store. We had one model that served as a baseline, which is like the basic t-shirt that everyone owns. It gave us a starting point for comparison.

Other models, like Decision Tree and Random Forest, are a bit more complex. They work by dividing data into smaller and smaller parts to make predictions. Using these methods, we could figure out what Factors were most important in predicting damage from earthquakes.

#Our Best Picks

After testing different models, we found that the Random Forest model worked best. Why? Because it combines predictions from many smaller models (like a bunch of friends giving you advice) and comes up with a stronger answer. This helps us reduce the chances of making wrong predictions.

#What Factors Matter Most?

When we looked at the results, we found that some factors mattered more than others. For predicting how many people might die, the earthquake's strength was super important. But when we looked at the number of deaths relative to the population, the number of people living in the area had the biggest impact. This makes sense; if more people live close to the epicenter of an earthquake, the chances of casualties are higher.

#The Bigger Picture

In addition to our predictions, we wanted to understand why things turn out the way they do after an earthquake. Sometimes the damage caused by fires or other disasters after the shaking can be just as bad as the earthquake itself. This isn't something we can accurately predict, but it shows how complex this topic can be.

Even though our model is helpful, there are still many things we don’t know. For example, we only looked at earthquakes that happened before 1950, meaning some of our data might be incomplete or inaccurate. We had a limited amount of information to work with and that can also affect what we’re trying to predict.

#Future Improvements

In the future, we plan to merge our findings with advanced earthquake prediction systems already in place. By bringing together different models, we could create a much more useful tool for disaster agencies in Turkey. They could better prepare for earthquakes, saving lives and reducing damage.

We also recognize there are limits to our research, and we want to keep improving it. As technology gets better, we hope to develop even more effective ways to predict damage from earthquakes.

#Related Work

The world of earthquake prediction has been growing thanks to technology. Newer models have been able to give us more information about where and when earthquakes might happen. Some organizations even host competitions to see who can best predict the damage caused by earthquakes! However, many of these models focus on building damage instead of human lives, which we think is very important to consider.

#Final Thoughts

Our work focused on predicting how severe earthquake damage could be in Turkey based on factors like building strength and how deep the earthquake strikes. We believe that understanding these factors will help improve how prepared people are when an earthquake happens. Our hope is that by sharing this research, more efforts will go into preventing loss of life when the earth decides to shake.

So, next time you think about earthquakes, remember that there are many smart folks working hard to keep everyone safe, using a mix of data, technology, and a little bit of luck. And let’s hope the ground stays calm!