IE11 Not Supported

For optimal browsing, we recommend Chrome, Firefox or Safari browsers.

Public Health Insights: Addressing Vaccine Distribution Through Sentiment Analytics

As the nation has worked to drive successful COVID-19 vaccination campaigns, a key factor to success has been understanding who remains unvaccinated and why.

Google 2-1-1.PNG

Improving Public Health Equity and Outcomes During Times of Crisis

Google Cloud’s IVIs was introduced in spring 2021. During the rollout, many states didn’t account for gaps in equity.

“For us, that was an opportunity to use sentiment to help the state pivot to an approach that was informed by the attitudes of Californians toward the vaccine,” said JP Petrucione, deputy director for communications and insights at California’s Office of Digital Innovation.

Petrucione said California officials developed several principles to guide response, including the simple but effective decision to launch feedback surveys on state websites to gather data in real time and at scale. This served as the basis of data to further combine with the other touchpoints the state had with its constituents, from web analytics, anonymized call center transcripts, social media and advertising data, Google Search trends and more.

Jeffrey Marino joined a data insights working group within California’s ODI in spring 2020 and was part of these survey efforts. Composed of volunteers, Google employees and individuals from the private sector, the group initially helped assess the impact of the governor’s stay-at-home order and whether it was moving the needle on public health.

“Understanding where Californians were going, whether or not they were staying home, and also their attitudes and beliefs in response to what the state was asking of them was a real, significant challenge in terms of data collection,” Marino, who serves as insights program manager, recalled. The group turned to the state’s websites, which were already experiencing a high level of constituent interaction, and leveraged Survey Monkey to ask questions such as, “In the last week, how many times did you visit a bar or restaurant?”

“We took the lessons that we learned from that experience and applied those to the vaccine efforts,” Marino said.

The group used a standardized and rigorous methodology to validate its data against other sources and to rule out selection bias. Visitors to were much more likely to proactively seek out information and want to get vaccinated than those using other California government websites.

To glean actionable insights into the over 80,000 qualitative responses, California officials leveraged a range of resources. Google Cloud BigQuery and Syntasa were pulled in, using Google Cloud Natural Language AI processing tools, geographic identifiers and data from the Healthy Places Index, which ranks socio-economic indicators of health to gauge where a community is on the spectrum of socioeconomic advantage.

Additionally, data from, California’s COVID-19 vaccine appointment website platform, was joined with administrative data around vaccines to answer questions about vaccine hesitancy and where vaccines were actually being administered.

“We were able to create a very actionable tool for state policy or decision makers around vaccine distribution,” Petrucione said. “One of the key things that we learned in this entire process was that in order to have an equitable distribution of any state resources, you have to meet Californians where they are.”

The data prompted the state to stand up mobile and pop-up vaccination clinics targeted at zip codes where vaccine hesitancy was higher. It also indicated where residents wanted to become vaccinated but couldn’t because they faced barriers to getting to traditional vaccine sites.

It became clear to us that sentiment analytics was not just a tool for understanding the tastes of Californians, but also it was a tool for measuring demand for a service in real time.

Leveraging Technology to Better Understand the Community

Looking to the future, Petrucione anticipates being able to replicate the process of gathering data for vaccine distribution and apply it to other crises and challenges —in the State of California and beyond.

“This is very much an experiment and we’ve been really stoked with the results of the vaccine sentiment work that we’ve done,” he explained. “And we see that that is really a proof point for whether sentiment — as we understand it and are using it, how we gather it, how we analyze it, how we apply it — can be replicable and scalable.”

Learn more about how Google Cloud can help your agency use sentiment analysis to glean valuable public health insights.
Google is a trusted technology partner who understands how to help agencies transition from legacy architectures and utilize their data to fuel true mission success. Google Cloud Platform provides cloud-native infrastructure with layered security, machine learning and analytics at web-scale to rapidly innovate and advance agency goals.