10 Lessons I’ve Learned Working as a Healthcare Data Analyst

The motto is — learn, explore, experience, and share

Hi! I am Rashi and I work as a Data Analytics & Reporting Consultant with a leading health insurance company based out of Chicago. Before starting my first full-time job, I interned with PepsiCo for about almost a year while in school, and from the time I became familiar with the word data & analytics, I have only been learning every day.

No two days are the same working with healthcare data and I love that. Every day puts you through a good challenge. It would now be almost two years since I started working full-time and I have gathered experience & maturity as a data professional, established best practices for myself and my team, and more importantly, learned from the mistakes that data analysts can be vulnerable to.

From collaborating with cross-functional teams and peers with years of experience with data and the organization, here are my ten learnings, in no particular order, that shall only evolve me further and I wanted to share those with my Medium community.

1. Empower your team & be a visionary

It is not only a manager’s job to keep the team spirit high. You must look out for your teammates and peers. For starters, ask for their ideas and insights, reinforce them with positive feedback and demonstrate that you appreciate their collaboration.

The biggest learning I had in the past few years is — if you have a seat on the table, make sure you speak up.

As a data professional, you always should stay in the loop with the industry advancements and suggest new avenues of growth and development for the team if there is anything that connects and is feasible.

The second realization, invest time in fostering relationships across the enterprise with people interested in driving innovation. In short, see your future for yourself and the organization and develop your vision accordingly.

2. Teamwork and networking open so many doors

Get those coffee chats on your calendar, attend community events at work, and don’t skip that happy hour with your teammates.

Over the last year, I made my life easy by the day by recognizing resources and peers with the knowledge I could leverage for my day-to-day job. Reaching out to the exact people who have the right answers for you is a total game-changer. But before that, you need to put in the work to network with folks across your organization.

Your teammates could have immense knowledge about what you are researching from a previous role but you would never know that if you did not connect with them beyond the good mornings and good nights.

I’ve learned that it is considerably easy to switch jobs when you already know someone from the hiring team. You know what they work on, and it is also easier to get poached for opportunities you wouldn’t explore otherwise!

Your network is your net worth.

3. Move from data projects to data products

Now, this is a mindset change I am talking about.

As organizations continue to make the big move to promote “data-driven decision making” across boards, a switch from a project-centric mindset to a product-centric mindset is core to this transformation. To understand the difference between a data project and a product, let me clarify how I perceive the two.

A data project could be temporary; you work on deliverables within clearly defined efforts in time and money and once the project is complete, you may move on. Now, a data product could be a tool or a solution designed to address a problem and the product can live through multiple stages until its value is realized.

Projects focus on completing tasks whereas products maximize user value

4. Power through the data lifecycle

There are countless interpretations of the term data lifecycle and below is my take on the lifecycle that has worked for me in my experience so far.

You work through — Business Understanding // Data Collection // Data Mining // Data Exploration // Feature Engineering // Predictive Modelling // Data Visualization and back to the business

Depending on your project, the steps you have to work on or not may differ, but laying out your project with the data lifecycle in mind has always steered me away from confusion and chaos mid-way.

5. Provide actionable insights from data (analyses)

It is incredibly important in today’s day and age to translate your data, analyses, and insights into a language that the business can understand — the business does not care what the accuracy of your model is. The executives are all about data-driven decision-making. For that, as data professionals, our key role is to provide actionable insights. I was extremely fortunate to learn this aspect of a data analyst early from my internship with PepsiCo.

Last year, I had an opportunity to interview candidates for a data analyst role and I created an assessment to test the candidates on the role's necessary skills. We asked them to create a dashboard talking about insights from their analyses. Only 2/10 candidates we interviewed could provide “actionable insights”. There’s a dire need for data professionals who can do that.

A thought for food, your audience sometimes has different skills or grasping power than you do. Therefore, you have to make the data & insights palatable for a non-technical audience.

6. Pivot to storytelling with data

At organizations, it is a standard practice now to present a ‘story’ — a narrative constructed from datasets put together to present broader implications. The sooner you learn storytelling with data, the easier it gets to understand what the business needs and demands and to interface with the executives (make people your front-row cheerleaders).

Storytelling with data not only translates data into insights, but also brings these insights together with qualitative and data analyses, statistics, visualizations, and domain expertise to better understand the business goal or problem.

Data stories can be easily mixed up with data visualizations. However, data visualizations are just a subset of data stories and are used as a tool to paint a larger picture.

7. Your hard skills will only grow by the day

When I started working, I took up a few online courses to sharpen my SQL and Python skills based on my everyday tasks to simply make myself more efficient. Over time, I have realized that you do not need to hone the skills you already know. You only get better with time. However, you can invest time in expanding your portfolio.

As data professionals, whether we address it or not, there are new skills that continue to erupt in the industry. Seeing new grads come with knowledge of a bunch of new tools and having to keep up with the growing expectations from a data professional is tough.

The way I see skill expansion is not learning supply chain management because it is hot. I have a career path in my mind (say the next 5 years from now) and work in that direction. If I want to be a data product manager in the next few years, what skills could lead me to the next promotion?

8. You have to automate everyday jobs

Automation is right at the door to everything you do each day.

Over the past few months, as I talked to more folks in data, I recognized that people who worked on monotonous tasks moved away from creativity, disrupting the productivity flow. Automation in your workflow has helped me bring out the best of me and the projects I work on, making it easier to manage the data more efficiently.

Imagine what you could do with all the additional time you get from automating your everyday jobs, investing your efforts in new analyses, and discovering new insights.

One tip with automating everyday jobs: start small (one task at a time)

9. Create documentation for everything you do

The biggest mistake that we do without realizing it until the boat is sinking is, not documenting what we do.

I would be working on a project for a year, using the same four data sources each day, I know the data dictionary, in and out even at 2 AM and there will be a day when I no longer work on the project. One year and nine months later, if I am asked to transfer the analysis to a new hire to put together an enhanced version, I am mad scrambling for access, assumptions created for research, and much more that I missed noting down.

Everyone in my data network at work has talked about the documentation of your everyday job and the value it adds to your professionalism. You have to get better at version control, sharing data and reports securely, commenting lines of codes, and documenting experimental methods that you rejected before deciding your final way to the analyses.

10. It is easy to debug broken pipelines than to prevent one from breaking

In my job from 9–5, I have to spend 60% of my time on days to maintain what is already functioning well to function well tomorrow, the day after, and a week later.

Data pipelines, architecture, models, and anything in the data lifecycle you work on could break at some point. But now, atleast you know what is to be fixed. When things are not broken, it is a constant effort to keep the jobs and models up and running every single day. This concern made it essential for me to identify what could break my pipelines and minimize the risk by substituting the tool, technology, or introducing automation.

That’s it from my end on this long blog post. Thank you for reading! I hope you found it an interesting read. Let me know in the comments about your experience with storytelling, your journey in data, and what you are looking for in 2023!

If you enjoy reading stories like these, consider signing up to become a Medium member from this link.

Happy Data Tenting!

Rashi is a data wiz from Chicago who loves to analyze data and create data stories to communicate insights. She’s a full-time healthcare data analyst and blogs about data on weekends with a cup of hot chocolate…