What happens when you marry social media and big data?
Splunk4Good pairs the two to target help after disasters
- By Mark Hoover
- Feb 15, 2013
There’s been a lot of buzz around how companies are using social media these days.
Most commonly, companies are using it as a tool to achieve greater efficiency and communication among its employees. One company is doing it a little differently, though.
Splunk is combining social media with big data analytics, and has developed a way to better respond to natural disasters, such as Superstorm Sandy, when it comes to where relief efforts need to be.
The company opened its federal practice around 2006, and now has more than 200 federal civilian, defense and intelligence agencies customers as well as state and local governments.
The company typically helps the government with network security operations and compliance, said Stephanie Davidson, director of Splunk’s federal civilian team.
But the government also relies on Splunk for IT operational intelligence, which looks at different data sources, and allows users to use web analytics, as well as manage and monitor different data.
This is where social media comes in, along with Splunk4Good; Splunk4Good was developed by Christy Wilson, vice president, product operations for Splunk, and is the company’s corporate social responsibility program.
Splunk4Good’s involvement with social media and analytics began when FEMA created an innovation team to figure out how to better respond to disasters such as Superstorm Sandy.
The team includes a number of different agencies and companies, like the Health and Human Services Department, Executive Office of the President, the non-profit sector, the private sector and one company in particular, Geeks Without Bounds.
Geeks Without Bounds, or GWOB.org, is a humanitarian accelerator that partnered with Splunk4Good, and were the ones who asked Splunk4Good to be involved in this process.
Once involved, Splunk4Good realized that many of the Superstorm Sandy respondents actively used social media during the storm, but for the purposes of outreach, rumor control, and things along those lines, Wilson said.
“They weren’t analyzing that data in real-time and making operational decisions based on it,” she said.
Splunk4Good showed them that you can take this data, which lends a micro-level perspective to the mix, and use it to make decisions about disaster relief, she added.
Recently, at FEMA’s Think Tank meeting on Innovative Solutions in Emergency Management, a FEMA employee who was not associated with the agency’s innovation team said that this kind of big data analytics would have been very helpful for him, Wilson said.
At one point, this person had to deliver over a million gallons of water in a single day; however, he had to rely on his own educated guess as to where it should be delivered.
Splunk4Good’s real-time analysis capability, combined with social media, would have helped him figure out where the need was and what was needed.
One example of how social media and big data analytics can work together involves Twitter and Instagram.
The company was able to go through Twitter and “splunk” it, pulling out any Instagram picture that included a hashtag or word like “Sandy” or “hurricane,” Wilson said.
These photos provide a ground-level view of the damage, providing photos that people can view and rate, on a scale of one to ten, in terms of the damage level, Wilson said. Based on those ratings, relief workers can prioritize areas that might need more assistance than others.
Another example includes charts with information on how many tweets were received on what day. The tweets were categorized by what they called for and helped map which resources were needed the most in a given area.
Looking at those examples, you see how “if you combine Splunk’s IT data with social media, and join that with your mission, then you can get some really interesting analytics,” Davidson said.
Mark Hoover is a senior staff writer with Washington Technology. You can contact him at email@example.com, or connect with him on Twitter at @mhooverWT.