Climate change is a very pressing issue that if not addressed soon could have catastrophic consequences in our future

According to 2019 Porter Novelli/Cone Gen Z Purpose Study 90% of people in the Gen Z Generation believe companies must drive action on social and environmental issues

Gen Z will make decisions on what brands to buy and use based on how a company is helping tackle the issue of climate change

So we thought what if there was a measurable way to find out how much a company is investing into combating climate change and know what actions they are taking?

Currently there is not a place where you can quickly search for company and find out exactly what actions they are taking to help combat climate change and provide a measurable metric to easily indicate which companies are sustainable.

What it does

WATWE is a search engine that provides a calculated unbiased total WATWE score and environmental headlines for any company the user asks for.


The total WATWE score is calculated by doing an internet search for articles about the company using the following five environmental issues as search keywords:

Water (ex. clean water access, waste water management)

Agriculture (ex. agriculture pollution, green farming, food security)

Transportation (ex. green transportation technology, reducing emissions, public transportation infrastructure)

Waste (ex. recycling management, composting)

Energy (ex. renewable energy, energy consumption reduction technology, clean air)

WATWE Headlines

WATWE uses the following news sources in its calculation:

Google News

Business Insider




Fox News

ABC News

Tech Crunch

The Verge

How it works

WATWE uses the first 5 articles from each category search and sends the text over to IBM Watson to calculate how positive an article is. The Total WATWE Score is the sum of all of the tone calculations IBM Watson provides.

How we built it

Amazon Web Services

Amazon API Gateway

AWS Lambda Functions (JavaScript and Python)


Alexa Custom Skill with APL Integration

Google Custom Search

IBM Watson





Challenges we ran into

Challenge: NewsPaper3K API unable to read text from non news websites Solution: Use Google Restricted API and limited the search to just a few news sources

Challenge: API Gateway has a max timeout of 29 seconds Solution: Created a search cache with DynamoDB so that Get Request does not do a full search for companies that were already previously searched

Challenge: Alexa Response has a max timeout of 8 seconds Solution: Created an additional lambda function so that Post Request can respond very quickly and still have the creation in progress without waiting for a response back

Accomplishments that we're proud of

Building a fully functional Alexa App with APL Integration that uses a ton of cool technologies including Amazon Web Services, Google Custom Search, and IBM Watson.

The fact we were able to finish a phase 1 of the application that literally does a google search for 25 articles, takes the text the from those articles, and has IBM apply tone analysis on each of those articles in rapid time is a huge accomplishment.

What we learned

We had very little experience with Amazon Web Services and no experience with Alexa Development, IBM Watson or Google Custom Search so this was a complete learning experience using technologies that we had little to no experience with which made completing the project all the more rewarding.

What's next for WATWE - The Green Search Engine

  1. Train a Custom Machine Learning Model to improve accuracy in search and score calculation
  2. Ability to authenticate and keep history of searches
  3. Ability to export searches into a PDF Report
  4. Compare multiple companies/multiple searches at once
  5. Add last search date and refresh a record if the last search was over a week
  6. Build a daily leaderboard to display the top company scores
  7. Ability to donate to environmental charities
  8. More variety of sources and methods for calculating the WATWE score like the company's website/blog and social media
Share this project: