Getting started on Zindi

• Reach out to Darius a mentor at START if you need any help
• Enroll in your first competition
• Create a team

The core of our challenge is to design and shape the future of landslide prevention and management with the example of Hong Kong.

Find all your information on the START Hack Git Repo here: https://github.com/START-Hack/UNEP-STARTHACK22

Here you can find a great article from BBC why this is an urgent topic!

Hong Kong, one of the hilliest and most densely populated cities in the world, is frequently hit by extreme rainfall and is therefore highly susceptible to rain-induced landslides. A landslide is the movement of masses of rock, debris, or earth down a slope and can result in significant loss of life and property. A high-quality landslide inventory is essential not only for landslide hazard and risk analysis but also for supporting agency decisions on landslide hazard mitigation and prevention.

As the common practice is visual, labour-intensive inspection, this hack focuses on automating landslide identification using artificial intelligence techniques and embedding this solution into the creative vision: “Harnessing the power of modern technologies to address landslide-management in Hong Kong”

During the hack, you can submit your predictions from your code with real-time ranking on Zindi, but only the final one will count and has to be sent together with the pitch to dominik.bisslich@un.org before 25 March 10:00 CET. We also ask you for two draft submissions on the 24 March with the code and pitch, the first at 10:00 CET and 22:00 CET. It doesn't have to be perfect, we would like to track your progress, guide you and support you as best as possible.

Our vision is to rethink landslide management in Hong Kong and integrate and make sense of modern technologies. Not only is quantification important, but also an entrepreneurial and innovative approach. The final product is therefore a vision of landslide management presented in 5 minutes based on classified input parameters using machine learning/deep learning. After that we got a short 3 min Q&A-session with our experts.

Building on our shared vision, the first step is to analyse the provided dataset and perform binary classification of landslides. Which input parameters such as precipitation intensity or slopes are more likely to cause landslides? For this purpose, we provide one dataset for training and one for testing. The choice of approach is up to you, whether machine learning or deep learning. We want to identify the variables that make landslides possible and use them to identify risk areas. Because of the wide range of possible analyses and influences, this task can be performed by people of all experience levels, from novices to those with extensive coding training. Machine learning knowledge is desirable, but not essential. The evaluation of this task runs on a F1 score, the highest wins this task.

Data, however, has no meaning without context. It remains questionable how these can be used. To make our vision become true we also need creative entrepreneurs that have the ability to innovate and rethink without boundaries. For example, can other data be used, can networking with other emergency services be made possible, can this format be transferred to other countries or cities? We deliberately did not give a focus here, but simply addressed "Harnessing the power of modern technologies."

Landslide management is already a big issue in Hong Kong, but the use of AI can help to transform it. There are several research projects about that right now. The dataset was created by Hong Kong University of Science and Technology.

We will evaluate your solution based on the technical sophistication and entrepreneurial spirit surrounding it. One tip for evaluating your technical model is using Zindi for real-time ranking and the current F1-Score of your submission. So you know whether to work more on the model or the creative solutions you are envisioning.

You may use R or Python to code your solution, we recommend using Google Colab as it allows access to GPUs.

Your submission should look like (where 1 indicates positive and 0 indicates negative):

ID label

00OADRP 1

012YMY8 0

014E83I 1

You will find the detailed criteria down here in the table:

The winning team will get four DJI Mini 2 drones with which landscapes or people are photographed from above. Perhaps there are also landslides included ;D

Since you always learn something in a challenge and it will be hard to pick a winner, the runner-up team will also get something. A hackathon always brings you to your limits, you cross them and this can go to your own health, we would like to promote your wellbeing after the hackathon. That's why we are adding four fitness watches. This is the Huawei Band 6, four of them in black. Enjoy and celebrate yourself.

All our Hack Mentors except for Haojie will be on Discord and present during the Hack. In the first three hours of the Hack, before the sample submissions and the four final hours we will be available at our table at the hack.

You can always contact us or our experts. If someone is attending online, we stated that in the description. Otherwise we are happy to set up a meeting and guarantee for changing roles at our table so you have a chance to meet everyone.

Feel free to reach out!

Meet your Hack Mentors

Meet Darius (Zindi Data Scientist)

Meet Melissa (Graduate student in Computer Science)

Meet Maxime (Graduate student in Robotics)

Meet Dominik (Graduate student in Mechanical Engineering and Business Administration)

Meet Haojie (PhD Geoscience at Hong Kong University of Science and Technology) via https://cehjwang.github.io/: Deep-Dive by HKUST https://raw.githubusercontent.com/START-Hack/UNEP-STARTHACK22/main/Media/Technical%20Introduction_By%20Prof%20Limin%20Zhang%20of%20HKUST.mp4?token=GHSAT0AAAAAABSTH36TGKSREO2AH7MZPINIYRYRWYA

Meet our experts in Disaster Management, Law, Emergency Response, and Project Management

Meet Muralee (Head of the branch Crisis Management Branch at UNEP)

Meet Paula (Project Coordinator for Modern Technologies at UNEP)

Meet Arielle (Master degree in Technology Law)

Meet Eike (Graduate student in Environmental Policy and Law)

Meet Abhijith (LLB Law; Chess Start-Up Founder)

Meet Matheus (Master degree in International Law)

Meet Aman (Graduate student in Geoscience)

Meet Thomas (Undergraduate student in Scientific Computing and Data Analysis)

Please follow the rules of START hack.

On Zindi you can participate as an individual or as a team of up to 4 members.

If you are not registered for START hack you cannot participate in this hackathon.

Submissions and winning

You may make a maximum of 75 submissions per day.

You may make a maximum of 225 submissions for this competition.

Before the end of the competition you need to choose 2 submissions to be judged on for the private leaderboard. If you do not make a selection your 2 best public leaderboard submissions will be used to score on the private leaderboard.

During the competition, your best public score will be displayed regardless of the submissions you have selected. When the competition closes your best private score out of the 2 selected submissions will be displayed.

Zindi maintains a public leaderboard and a private leaderboard for each competition. The Public Leaderboard includes approximately 40% of the test dataset. While the competition is open, the Public Leaderboard will rank the submitted solutions by the accuracy score they achieve. Upon close of the competition, the Private Leaderboard, which covers the other 60% of the test dataset, will be made public and will constitute the final ranking for the competition.

Note that to count, your submission must first pass processing. If your submission fails during the processing step, it will not be counted and not receive a score; nor will it count against your daily submission limit. If you encounter problems with your submission file, your best course of action is to ask for advice on the Competition’s discussion forum.