Estimated reading time: 3 minutes
Do you how much it costs for communities to rebuild after natural disasters like hurricanes, tornadoes, and wildfires? According to Munich RE, 2018 was the fourth costliest year for natural disasters in U.S. history, totaling more than $160 billion in damage. The extreme weather that hit last year also took hundreds of lives and destroyed tens of thousands of homes and livelihoods.
Hurricane Maria hit Puerto Rico in 2017 and was one of the most destructive natural disasters in recent memory. Not only did it take 11 months to fully restore power to the island, it also took more than a year for government officials to announce the official death toll, which amounted to more than 2,000 people.
At Stratifyd, our goal is to use AI and machine learning to answer the questions nonprofits have been asking for years: How can we provide targeted aid after a disaster? How can we quickly reunite families who were separated from one another? How can we accurately measure the extent of damage caused by a natural disaster and help communities begin to rebuild? How can we rescue people before it’s too late?
As technology continues to advance and new tools are created, NPOs must face disasters head-on with analytics software solutions like Stratifyd, or death tolls will be higher.
Just like for-profit companies, NPOs need to utilize AI and machine learning to analyze the voice of the consumer (or those in need, in this case) to provide better aid when disaster strikes. NPOs can gather real-time insights about what individuals need in times of crisis based on geographical trends or verbatim feedback. Having these actionable insights pertaining to those in need could help provide targeted aid like medical supplies, food, water, and shelter to the specific locations that are in desperate need.
We’ve created a dashboard demo to showcase how our AI-powered platform can help NPOs paint a clear picture from the voices of those in need.
Analyzing 97,000 #HurricaneMichael Tweets in Minutes
We ingested tweets associated with #HurricaneMichael to analyze the voice of those in need after the initial impact. While 85% of the data occurred during the peak of the storm, the remaining 15% contained crucial information about the aftermath of the disaster. This smaller dataset tends to be overlooked in real-life situations but contains cries for help from people stranded in their homes, in need of resources, or not receiving national aide. Stratifyd applies sentiment to these important voices to bring them to the forefront before it’s too late to act.
Of the data we ingested, 21% contained bigrams or statistically relevant word clusters about topics like water distribution, shelter, and official response. By dissecting the data to focus on these trending topics, we can filter out unrelated topics and reveal thousands of records with negative sentiment that highlight the needs of specific areas on a heat map.
Breaking Down 50,000 Data Points from the West Coast Wildfires
Stratifyd ingested over 50,000 data records from news articles, forums, blogs, and social data about the California wildfires that tore through Sacramento and Los Angeles. Thousands of articles focusing on awareness, evacuations, and resources were posted by local and national media. Our neural sentiment model applies machine learning to unstructured data, taking human bias out of the equation while searching through a massive volume of data. As a result, we uncovered 125 records out of 50,000 that had negative sentiment attached to "camp fire," "utility company," and "hold accountable." These hidden clues multiplied at a 100x or even 1000x scale puts into perspective the utmost importance of small-scale data, and the effect it can have on saving lives or preventing future disasters.
Can your company or organization use our AI for good? Schedule a demo today to discover how we can solve the world’s biggest problems together.
Thank you for subscribing!
You will be the first to get our new blogs. Stay tuned.