Predict Wildfire Intensity with Watson Studio Machine Learning

For this Code Pattern, we will use data on Wildfires from NASA to predict the intensity of Wildfires, using Watson Studio Machine Learning. NASA provides data for various things, from weather and climate to solar flares and wildfire. This data is paid for by U.S. taxpayers and is free to use. The missing component is machine learning, which can take data and train a model to predict one of the features of the data set. For this example, we'll grab wildfire data and build a model that can predict intensity of the fire base on latitude and longitude.

When the reader has completed this Code Pattern, they will understand how to:

• Use Watson Studio Machine Learning to train a model.
• Gather data from NASA for Wildfires.
• Create a predictor for intensity of wildfires based on latitude and longitude.
• Create a Node.js server that can utilize the Watson Machine Learning service for predicting wildfire intensity.

Flow

1. Use Watson Studio to add data assets and services.
2. Create the Machine Learning Model in Watson Machine Learning.
3. User interacts with Web UI to choose location of a fire.
4. Web UI interacts with ML model to predict brightness of fire.

Included components

• Watson Studio
• Machine Learning

Featured technologies

• Artificial Intelligence: Artificial intelligence can be applied to disparate solution spaces to deliver disruptive technologies.
• Node.js: An asynchronous event driven JavaScript runtime, designed to build scalable applications.

Prerequisites

• Install Node.js and npm (npm version 4.5.0 or higher)
• For this pattern you will need to download current wildfire data in CSV format. Please choose 7d for 7 days worth of information.
• You may additionally download archived data which will take time due to the request process.
• You will need a Maps API key.

Steps

1. Clone the repo

2. Create Watson Studio services with IBM Cloud

3. Get the credentials

4. [Either Deploy to IBM Cloud or Deploy locally]

   3a. Deploy to IBM Cloud

   3b. Deploy locally

1. Clone the repo

Clone the predict-wildfires repository locally. In a terminal, run:

$ git clone https://github.com/IBM/predict-wildfires

2. Create Watson Studio services with IBM Cloud

• On Watson Studio choose New project:

and pick Data Science:

• Name the project and select your object storage:

• Choose + Add to project -> Data asset or click the 10/01 button and add the data you downloaded in Prerequisites (MODIS_C6_*_7d.csv):

• Under the Settings tab -> Associated Services choose + Add service -> Watson and select Machine Learning. Either choose an Existing ML Service or create a New one.

• Under the Settings tab -> Associated Services choose + Add service -> Spark. Either choose an Existing Spark Service or create a New one.

• Under the Assets tab -> Models choose + New Watson Machine Learning model. Pick a name, the default Select model type of Model builder, and Automatic for "Prepare my data...".

• For Select Data choose the MODIS_6_\*_7d.csv file you added. When finished click Next.

• For Select a Technique -> Column value to predict (Label Col) choose brightness. This represents the intensity of the fire as measured by the brightness detected from a satellite.

• For Feature columns choose the features: latitude and longitude.

• For a Technique choose Linear Regression and Click Next.

• The model will train, and then Save the model.

• If you have downloaded archived data, you may create another model using that data using the same process.

• Once you've created the model, go to Deployments and click +add Deployment. Name it, keep the default Web service, and click Save. Leave this page open for the next step...

Get the credentials

• If you haven't left it open from the previous step, from the Watson Studio project page, under the Assets tab, double click the Watson Machine Learning model you created earlier. Go to the Deployments tab and double click to open it.

• Go to the Implementation tab and copy the Scoring End-point at the top. You will use this as SCORING_URL in either the IBM Cloud Runtime Environment Variable or the locally deployed .env file.

• On your Watson Studio project page, got to the Services tab on the top menu bar, and navigate to the Watson Machine Learning service you created earlier. Double-click the service, go to Service Credentials -> View Credentials.

• You will need your Maps API key.

• You will use the username, password, url, and 'maps API key' in either the Runtime Environment variables when running on IBM Cloud or in the server/.env file when running locally. These will be WML_URL, WML_USERNAME, WML_PASSWORD, and MAP_APIKEY.

3.a Deploy to IBM Cloud

• Press the above Deploy to IBM Cloud button, under IBM Cloud API Key: choose Create+, and then click on Deploy.

To monitor the deployment, in Toolchains click on Delivery Pipeline and view the logs while the apps is being deployed.

• Once the app has deployed, Click on Runtime on the menu and navigate to the Environment variables tab.

• Update the 5 environment variables with the SCORING_URL, WML_URL, WML_USERNAME, WML_PASSWORD, and MAP_APIKEY that you saved during Get the credentials. The app will automatically restart and be ready for use.

3.b Deploy locally

• In the server/ directory, move the env.sample file to .env.

• Populate the file with the SCORING_URL, WML_URL, WML_USERNAME, WML_PASSWORD, and MAP_APIKEY that you saved during Get the credentials.

SCORING_URL=""
WML_URL=""
WML_USERNAME=""
WML_PASSWORD=""
MAP_APIKEY=""

• Install the nodejs modules with npm install.

• Start the app with npm start. The app will now be running and available on you machine in a browser at http://localhost:3000.

Sample Output

Links

• IBM Call for CodeAnswer the Call for Code by building global solutions for disaster preparedness.
• NASA FIRMS (Fire Information for Resource Management)Fires in the last 24 hours.
• Can Scientists Predict a Wildfire's Spread?
• Predict a Wildfire's Behavior in Real Time
• Watson Node.js SDK

Learn more

• Artificial Intelligence Code Patterns: Enjoyed this Code Pattern? Check out our other AI Code Patterns.
• Data Analytics Code Patterns: Enjoyed this Code Pattern? Check out our other Data Analytics Code Patterns
• AI and Data Code Pattern Playlist: Bookmark our playlist with all of our Code Pattern videos
• With Watson: Want to take your Watson app to the next level? Looking to utilize Watson Brand assets? Join the With Watson program to leverage exclusive brand, marketing, and tech resources to amplify and accelerate your Watson embedded commercial solution.
• Data Science Experience: Master the art of data science with IBM's Data Science Experience
• Spark on IBM Cloud: Need a Spark cluster? Create up to 30 Spark executors on IBM Cloud with our Spark service

License

Apache 2.0