Parkinson’s UK takes battle for a cure to the cloud with Snowflake Computing

Profile picture for user jtwentyman By Jessica Twentyman November 7, 2018
Summary:
Director of Digital Transformation Julie Dodd believes that a joined-up data approach could hold the key to tackling the neurodegenerative illness once and for all.

JulieDodd
Julie Dodd

In recent weeks, a large-scale study was published in the journal Science Translational Medicine that suggests that Parkinson’s disease could originate in the appendix and subsequently travel to the brain.

Conducted by researchers at the Van Andel Research Institute in Michigan, the research focused on the health records of more than 1 million people in Sweden, and its findings suggest that having an appendectomy early in life is linked to a 19% reduced risk of developing the neurodegenerative illness, in which the loss of neurons in the area of the brain that controls movement leads to tremors and slurred speech in patients.

The response to this research from Parkinson’s UK, a charity that works for better care, treatments and quality of life for people with the condition, has been cautiously optimistic. As Claire Bale, the charity’s head of research puts it:

In most cases, the causes of Parkinson’s are a mystery. But understanding how the condition starts and progresses is the first step to stopping it. If we can couple this understanding with tests that detect the earliest changes and treatments that can stop it progressing, we will have a real pathway to preventing Parkinson’s.

Tireless campaigning

In the meantime, the charity continues to work tirelessly to help those who are already affected - and data plays an increasingly large part in that work, according to Director of Digital Transformation Julie Dodd, a former creative director at the BBC.

This is the thinking behind Parkinson’s UK’s recent deployment of a cloud-based data warehouse from Snowflake Computing, which is helping it to consolidate information previously held in a wide range of disparate legacy databases and organize it for better analysis and reporting. Says Dodd:

What we had before was the fairly classic picture of an organization that had accumulated growing datasets in various systems over time, but hadn’t ever really thought about those data assets strategically. The only was data was being integrated was manually and that was an arduous and painful process. That’s no way to run a modern organization. It’s not efficient, it’s really hard work and it’s prone to error.

What was needed was a modern data architecture that could really serve Parkinson’s UK’s business needs, and an extensive search was mounted, in which a number of approaches were considered and a number of proof of concepts carried out. But many of these options required expensive hardware or access to technical resources that a charity would struggle to justify, says Dodd. Many, she felt, would take too long to implement or might not prove scalable.

Eventually, Parkinson’s UK hit on Snowflake Computing, which appealed for its cloud-based approach. Not only that, but the team at Snowflake spoke a language that the team at the charity understood and appreciated:

They spoke like I feel modern technologists should do - thinking about our business challenges, appreciating our need for a lightweight architecture, considering how our organization and its data needs are likely to grow and scale over time.

Data needs

So what do those data needs look like? According to Dodd, it’s a question of better organising fundraising data, certainly, but over the longer term, combining that data with information gathered by the charity’s direct services:

So we run a helpline and we run a local advisor service of people who go out and work with those directly affected by Parkinson’s to help them cope on a day-to-day basis. The information collected there becomes a lot more powerful when we can integrate it with fundraising data.

Then we can see things like the person who gave us a £1,000 donation was a person who called our helpline six months ago and has since been in touch with a local advisor. We can understand more about what their real needs are and we can start proactively tailoring the support we provide.

Every organization wants a 360-degree view of its customers, but for us, it means we can genuinely improve the quality of support they receive so that they get to live well with Parkinson’s - something that’s fundamental to our mission.

On top of this, new data initiatives are underway at Parkinson’s UK that focus on how it can share relevant insights from its work with other research institutions in real time, with a view to speeding up advancements and communication about the condition, says Dodd:

So this is about using our data strategically to help the entire global research community to understand the disease better, to target specific treatments and hopefully find a cure. That involves working with pharmaceutical companies, universities, other charities and data-sharing is what will help us all work together to get faster results.

Progress to date

The first step in the project was to identify key systems and processes that would most immediately benefit from being migrated to Snowflake. While creating Snowflake warehouses was relatively simple, says Dodd, the bigger challenge lay in understanding and interpreting data from source systems and cleaning it:

There have been some learnings for us here and many of our third-party systems use text files for transferring data. One of the functions that Snowflake offers to us is the ability to manage JSON [a language-independent data format] and we’re now transitioning our source systems to use this format more.

Parkinson’s UK is still in the early stages of transferring data to Snowflake and still has a way to go in connecting all of its different data sources, she reports. But Snowflake’s technology has made some of the ‘harder’, technical jobs easy, so that she and her team can get on with the task of really understanding data and figuring out new ways of presenting it to users.

The charity is using SQL Server and Tableau for querying and reporting, but isn’t particularly wedded to these tools, she says:

There are many other alternatives and cloud licensing gives us the freedom to choose. But since Snowflake supports SQL for querying, we’ve been able to utilize existing skills without needing additional training. As we onboard more data sources, especially those with large volumes of unstructured data, we expect to be able to leverage Snowflake’s feature set even more.

These are all important steps, she says, along the road to finding a cure for Parkinson’s:

And that, after all, is what we’re really here for. I believe that a joined-up approach to data, along with cloud technologies, will really help us get there faster.