A tale of one organization: learning how to swim in the ocean of open source (Katerina Gapanenko)
A tale of one organization: learning how to swim in the ocean of open source Speaker(s): Katerina Gapanenko Abstract: The Canadian Institute for Health Information is diving into open source, transitioning our 300+ analysts and all our analytics code from proprietary software to R and Python. We started 3 years ago, and over that time we have migrated 700,000+ lines of code, navigating waves of change while adapting to new IT environments and workflows. Challenges like technical barriers, workload balancing, and resistance to change were tackled through training, mentorship, and executive support. More than tools, this shift transforms our culture—turning siloed developers into fearless swimmers. Success requires training, teamwork, and courage. Organizations making a similar leap should start small, invest in training, and stay adaptable to stay afloat. posit::conf(2025) Subscribe to posit::conf updates: https://posit.co/about/subscription-management/
image: thumbnail.jpg
Transcript#
This transcript was generated automatically and may contain errors.
Hi, it's wonderful to see you here. It's been a great conference with lots of deep and technical conversations But my presentation isn't technical. I just want you to have some fun and hear my story Every story got a lesson and mine isn't different. It has a few and I hope you enjoy them and find them useful
Imagine this You are a team of professional swimmers. You've been swimming in the swimming pool for years for 20 years The temperature has been comfortable. The water is calm. Each swimmer has their own lane it's warm bright and cozy and One day you are told that the pool will be closing forever You will have to learn how to swim in the open waters in the ocean that you've never seen before let alone swamming
Your imagination goes wild strong winds rough shores, scary creatures It's how we felt in my organization when we started that migration When we decided to move from proprietary toolkit to open source R and Python
My name is Katerina Gapanenko I work at the Canadian Institute for Health Information a national organization That's for 30 years collects analyzes and reports health information There I led a project of this massive transformation from proprietary toolkit to open source
we have over 350 analysts and we have tons of code and over 500 products and Some of the code has been developed 20 years ago by people who are not longer with the organization and we treat this code as Vintage wine. Today, I would like to share the story of our transformation It was a multi-year massive project with a lot of stories to tell
One presentation ago you heard Yvonne Kina's story She was talking about the ripples of change her story was about the same project as mine But I'd like to share my perspective as the business owner I will tell you everything from the beginning to the end from top to bottom and If you're considering a similar transformation in your organization, you may find both stories useful
Why we made the leap
So three years ago, we stood at the edge of a cliff the water looked powerful, but intimidating We couldn't just test the waters. We needed to jump so we did but why we jumped
first Sustainability proprietary toolkit has been our safe pool for years Stable vendor-supported but inflexible and Expensive and it was hard to integrate into existing Workflows and we were modernizing our workflows as well
Second the talent more and more analysts coming to the organization with existing R and Python knowledge They were frustrated with our outdated tools Our existing staff were curious about what's going on outside of these proprietary tools in the open waters
Innovation and collaboration We couldn't just keep up with the speed of a modern data science if we stayed in this closed ecosystem We felt that we were missing the boat in this one. And that is why we decided to go from closed tools from proprietary tools to open source
Three dangers and three aids
Very quickly way and discovered three things that could pull us under Technical hurdles we needed to build new environment new ideas Parallel computing get we needed to learn all about all these tools And we felt at that time that we see just a tip of the iceberg
Then the next one is cultural resistance Learning in general is emotional like waves. It comes with excitement frustration anxiety and joy Learning major skills at work is even more difficult As an organization we understood that this change will not be smooth
The workload we had a lot of product to deliver We had too little time to learn and we had no time to migrate the code The constant pressure Circled beneath us as sharks if we didn't address it, it could have been fatal
To help us overcome all these three dangers We built three aids at three different levels at the organizational level. We built executive support At the project level. We built a diverse project leading team a small team and At the individual level at the analyst level we built mentorship and training programs These three aids help us to survive and succeed without them. We might have sunk
Executive support
Executive support I see the executive support as a lighthouse as the beacon guiding everyone through the rough waters Steady and visible and given directions
We started gradually at the beginning. We plan to migrate 50% of our code in several years But it was very challenging to figure out how to split the organization in two How do we go about it? Is it gonna be right hand and left leg? Is it gonna be ways up and ways down how to decide who will stay in the closed? Environment and who will go and try the open open waters open source at the same time we started talking to a lot of external organizations who started the same transition as ours and All of them who maintain and maintain two systems at the same time told us that it was very challenging
Because you and they needed to maintain two servers They needed to maintain two groups of analysts and it was difficult to collaborate among analysts and share their code so we came back to our analysts and to our executive team and we explained to them all these challenges and With their support we decided to make one bold call
single hard deadline every script every data set and every analyst Our executive gave us space and cleared our workload They made learning a priority for everyone in the organization They showed their support and they were a part of our celebrations We created a few KPIs to tell about the progress in training immigration and in the environment adoption
The project leading team
Our second success factor was the project leading team itself Think about it as a rescue team with diverse strengths and skills and tools small versatile and ready for our waters and designed to save lives
We had different strengths and different skills some of us new hi-hi business Others new hi-hi products very well and the legacy code Others new are in Python, but some of us were not technical at all But they were able to read the room and manage emotions and they were able to organize all Sorts of events that we had throughout the projects Others were great with project management and reporting
We were just a handful of people with different strengths, but we were bound by the same mission But despite of our diversity we operated as one crew. We did everything together brainstorming making decision Executing and we did it fast because we trusted each other
And we build trust around us We instilled trust in analysts that they could learn and they could deliver We build trust with executives that our project team could deliver. We showed everyone that the project would succeed That mean being honest about challenging Listening to everyone's worries trying to unpack those of worries and get to the bottom of their concerns and show empathy
All this communication took time and patience in many cases we needed to adjust people as expectations And tell them in advance what's happening in some case in case Adjusting expectations wasn't easy. I will give just one example
At the very beginning our analyst felt that the new environment should be as stable as the old environment So in in in other words, they didn't expect the ocean water to be salty with seagrass and jellyfish So we had to explain to them in several different ways and several at several different occasions. That is not the case
But through this communication, we also understood their worries We understood what was the barrier for them and why stable communication stable environment was important for them So eventually we were we were adjusting their expectations, but we also Change our view on what they were worried about so we found the middle ground
The whole transformation wasn't just technical. It's also Emotional and it's the emotional in the first place Understanding acknowledging addressing people's worries is at most important
The whole transformation wasn't just technical. It's also Emotional and it's the emotional in the first place Understanding acknowledging addressing people's worries is at most important
Training and mentorship
Then the training when we realized how much code we need to migrate We started looking elsewhere to understand how we can deliver the project within the time frame that we set We explored hiring external gurus or hiring an army of co-op students Outsourcing the whole thing outside of our organization or using automation We waited all these options and decided to invest in our analyst in the people we already had We wanted to equip our analyst with the tools and skills and support they needed
We knew analysts couldn't just jump in the waters and start swimming so we built a learning ecosystem for them We created variety of options for everyone to choose from it was self-paced or in class short long Created inside the house or off-the-shelf courses at the beginning They benefited the most from our courses that we built ourselves Ourselves with our data and our use cases where we showed our code and converted it throughout the course
Later, they were comfortable With learning taking courses off the shelf and learning by themselves on their own Those who were scared to start we organized code alone and get together events Where in small groups people were sitting next to each other Actually in the same class with the support of mentor They were doing a project together a small little project but together that helped a lot
We also gave people GenAI tools and Not to write the code but to learn To learn the code that was created by somebody else who left the organization ages ago and forgot to leave any documentation whatsoever we they use these GenAI tools to Convert the existing code into pseudocode in plain English code to understand and better and to find efficiencies and also to To learn the new code that was created by somebody else in the team
One analyst called ChatGPT a patient mentor a tutor Available day and night to answer their questions that were honestly embarrassed to ask their neighbors Their the neighbors in the in in the office or their mentor
Altogether our analyst completed 40,000 hours of training they were Scared at the beginning and we were scared to give them that much training But at the end they appreciated the the challenge and one analyst said learning was fun I appreciated the opportunity it brought excitement to my day-to-day work
But the real game-changer was structured mentorship We hired a brand new team of mentors not based on their sharpest code, but based on their ability to teach and ability to create psychological safety
Every team got a mentor and every mentor supported several teams. They were providing technical support Answering questions help me sue the code with our in Python code with server questions They also helped build confidence in Analysts, they were talking to them individually and in group They comment them down if it's needed every team went through Emotional transformation along the project But the way the mentors support they felt safer their words not mine
The full picture
To see our project on one slide as one image think of our journey like being out in the open water There is a lighthouse the executive support that is steady visible and kept us oriented No matter how rough the waters got it cleared the workload and gave us space to learn
Then it's a small diverse and trusted crew Roaring and sink and providing support to everyone and building trust in everyone They listen they supported they encouraged they appreciated the effort
and Then the swimmers in the water Analysts equipped with training Mentors and the safety net to admit that they didn't know something put together That's how the transformation became possible
Where we are today
When we first stepped into the ocean of open source the waves looked big and the current felt strong But today we're not just staying afloat. We are swimming confidently together We made the leap by investing in our own analysts together
350 people migrated one and a half million lines of code over 500 products and over 18,000 data sets today hundred percent of our work is happening in R and Python We started sharing our code internally inside our organization We are donating pieces of our code in the common code repository our analysts our old culture isolated and cautious now Slowly, but surely has been changing to something more collaborative and more experimental and this is only the shoreline
the future The ocean hasn't gotten any smaller tech keeps evolving Open source keeps expanding health data science keeps raising ahead and that's okay Because now we are ready for it. We keep learning keep pushing and keep building in the future We can focus on speeding up our delivery and that will give us time to learn something new and to innovate
Today, we are not the same organization as before. We didn't just adopt open source We adopted openness We see ourselves Contributing to public open source projects now many of our clients and partners in Canada and abroad have started Exploring open source tools same as us. They started coming to us and asking for advice for our recipe for success So today we are not just swimming ourselves. We are guiding others through their own transformation We are making ripples beyond our organization
Today, we are not the same organization as before. We didn't just adopt open source We adopted openness
I will leave you with this if you're standing at your own cliff wondering whether to jump Don't wait for the waters to feel warm. If we could do it slow heavy legacy bound organization You can do it too. My advice jump in the water is fine
Behind this transformation was a team of people who made it happen We have more lessons to share so if you want to talk to us, here is how you can reach us. Thank you