AI for the
next disaster.

Tonight you build an AI prototype for the natural disasters our region faces: wildfires, floods, earthquakes, heat and drought. Everything you need is in the tabs below.

Open the disaster map Table guide (PDF)Set up OpenCode Field guide

Hack for Humanity, The AI Collective Jerusalem, in partnership with the Hebrew University of Jerusalem, ASPER-HUJI Innovate and The Medium

01Table guide

The worksheet from your table.

Joining online, or lost your copy? This is the same two-page guide every table has: the mission, the build plan, the data and how you're judged.

Open the table guide (PDF)Download PDF

Table guide, front: agenda, mission, problem statement and build plan

Page 1 · Mission and plan ↗

Table guide, back: data, prompts, AI toolkit and judging criteria

Page 2 · Data, tools and judging ↗

Fill it in online

Copy this into a shared doc or your team chat and fill in the blanks together. Your problem statement is also the first line of your demo.

team worksheet

Team name:
Challenge (Wildfires / Water crises / Air quality / Earthquakes / Droughts):
Phase (Predict / Prepare / Respond / Rescue / Recover):
One real user we help:

Problem statement:
For {who}, facing {what problem},
in the {phase} phase, we build {what}
so that {what changes for them}.

Build plan:
18:45  Pick and scope: challenge, phase, user, statement
19:05  Explore the data: filter the table, download the CSV
19:20  Build the smallest thing that works (Amazon Bedrock for the AI)
20:20  Freeze and rehearse: run the demo end to end twice

02Judging

How you're judged.

Six criteria, 1 to 5 points each, 30 in total. A prototype doesn't have to fully work to win: judges reward well-reasoned ideas. Ties go to Problem & Impact, then Prototype & Execution.

01 · 5 points
Problem & Impact
A real need in the crisis, with clear people who benefit.
02 · 5 points
Crisis Phase Fit
Clearly targets one phase (Predict, Prepare, Respond, Rescue or Recover) and fits how that phase really works.
03 · 5 points
Use of AI & Data
AI does real work, and data (the NDS dataset or your own) is used well.
04 · 5 points
Prototype & Execution
How far you got in two hours. A working demo is great; a clear mockup and walkthrough also scores.
05 · 5 points
Feasibility & Responsibility
Could it be piloted here? Local context, privacy, and what happens when the AI is wrong.
06 · 5 points
Pitch & Team
A clear 3-minute story, a good demo, strong answers in Q&A.

04Setup

OpenCode + Amazon Bedrock.

Each team gets Amazon Bedrock access for the night. The key and region are on the guide at your table. Prefer a no-code tool? That’s fine too: see the tools in the field guide.

Step 01

Install OpenCode

terminal

# macOS / Linux
curl -fsSL https://opencode.ai/install | bash

# or with npm (macOS, Linux, Windows)
npm install -g opencode-ai

# or with Homebrew
brew install anomalyco/tap/opencode

Step 02

Add your Bedrock key and start it

Prefer not to use the terminal for the key? Start opencode, type /connect, search for Amazon Bedrock and paste the key there.

macOS / Linux

# Paste your Bedrock API key and region (from your table guide)
export AWS_BEARER_TOKEN_BEDROCK="paste-your-key-here"
export AWS_REGION="us-east-1"

# Start OpenCode inside your project folder
mkdir my-project && cd my-project
opencode

Windows PowerShell

$env:AWS_BEARER_TOKEN_BEDROCK = "paste-your-key-here"
$env:AWS_REGION = "us-east-1"
mkdir my-project; cd my-project
opencode

Step 03

Prefer the browser? Open OpenCode web

Starts an OpenCode session in your browser instead of the terminal. Run it after setting the key, and keep the terminal open while you work.

terminal

# In the same terminal, inside your project folder
opencode web

Step 04

Pick a model and get going

inside opencode

/models      pick a model, e.g. a Claude Sonnet model on Amazon Bedrock
/init        let it read the project and write AGENTS.md
Tab          switch between Plan (thinks, no edits) and Build (edits files)

Optional: save the region in the projectShow

Put an opencode.json in the project root so everyone on the team uses the same region. Settings in this file win over environment variables. Keep the key itself out of it.

opencode.json

{
  "$schema": "https://opencode.ai/config.json",
  "provider": {
    "amazon-bedrock": {
      "options": { "region": "us-east-1" }
    }
  }
}

Keep the key secret. Never paste it into a prompt, a slide, a screenshot or a file you push to GitHub. If /models shows no Bedrock models, check the key and region first, then ask a mentor.

05Starter prompt

Your first prompt, filled in.

Download the data into your project folder, replace the parts in braces and paste it into OpenCode. More techniques in Prompts that get it right.

paste into opencode

Read data/data_dictionary.json, then load
data/israel_jerusalem_natural_disasters.csv.

I'm building a {Predict / Prepare / Respond / Rescue / Recover}
tool for {who}. Build a simple Streamlit app that:
1. {first thing it shows or does}
2. {the AI part: what the model reads and what it answers}

Keep it to one file. Plan first, wait for my OK, then build.
Tell me the exact command to run it.

The data

Explore the disaster data.

287 natural disasters in Israel and the Middle East, on a map or in a table you can filter and download as CSV.

Map view →Table view →

AI for the
next disaster.

The worksheet from your table.

How you're judged.

Pick a moment in the crisis lifecycle.

Your first hour.

287 disasters, mapped.

OpenCode + Amazon Bedrock.

Your first prompt, filled in.

Two minutes, one story.

The field guide, for any build.

AI for thenext disaster.

The worksheet from your table.

How you're judged.

Pick a moment in the crisis lifecycle.

Your first hour.

287 disasters, mapped.

OpenCode + Amazon Bedrock.

Your first prompt, filled in.

Two minutes, one story.

The field guide, for any build.

AI for the
next disaster.