High-Level Project Summary
We developed a landslide alert chatbot system including information from local community people. In the prediction of landslides, changes in the local environment are very important but they are currently rarely used. Therefore, we thought that by asking local people to provide such changes, we would be able to predict landslides with high precision in a short period of time. To collect information, we used "LINE", the most popular SNS in Japan. With this system, people can be helped by alerts. It can also provide useful information to government decision-makers. In the future, the development of a method to quantitatively evaluate local information will lead to more accurate predictions.
Link to Project "Demo"
Link to Final Project
Detailed Project Description
We developed a landslide alert system that includes information from local community people. Currently, landslide prediction is based on model simulation with field observations, machine learning with static and dynamic satellite data, and field surveys by experts. But there are some problems like space and time-resolution are limited, and surveys by experts are expensive.
In the prediction of landslides, changes in the surrounding environment (e.g., ground noise, new springs, and groundwater, etc.) are very important information but they are currently rarely used. Also, satellites cannot detect local information in a short period of time. Therefore, we thought that by asking local people to provide such changes, we would be able to predict landslides with high precision in a short period of time.
The purpose of our project is to predict the occurrence of landslides at an early stage based on information obtained from local people and to warn people in areas where landslides are expected to occur in danger.
We developed an interactive chatbot as a system to collect information from local people. The frontend interface is “LINE”, which is the most popular SNS in Japan, and the backend system is Flask with python. For the test environment, the backend system runs in “Google Colab” and is exposed to the public internet by “Ngrok”. We used our smartphone to provide the local information and get the alert through “LINE”.
This system will have the following impacts.
- People will be saved by the alert.
- To get useful information for the government to make a decision.
- This chatbot can provide information to the informational vulnerable (e.g., deaf, cannot go outside, etc.).
The following future work will lead to more accurate predictions.
- Develop a method for local information to qualify local information. e.g.) A system that predicts precipitation using video.
- Local information is used for the prediction model. e.g.) Local precipitation (from 1) is used for the SHALSTAB model.
- High-risk area in hazard map. e.g.) Use NASA’s “ASTER Global Digital Elevation Model” and/or JAXA’s “Today’s Earth rainfall”
Space Agency Data
The effect of our alert chatbot system will be enhanced by below space agency data.
- “ASTER Global Digital Elevation Model”
- JAXA’s “Today’s Earth rainfall”
Hackathon Journey
Two of our members are competing for the first time, and one for the third time.
We had a hard time because we were unfamiliar with the filed of this challenge and the format of space agency data.
References
Global Judging
This project has been submitted for consideration during the Judging process.