Showcase 3: Emerging Expertise in Data Storytelling
This showcase illustrates my development in a key career skill, data storytelling, using the examples of my graduate Statistics for Data Science class, Software Product Operations internship, and the MISO Xtern Data Science Challenge.
Introduction
During my undergraduate career, I have sought out many opportunities to improve my skills in data storytelling. Analysis and visualization of datasets have always been fascinating to me, and I have worked on many projects that allow me to go beyond theory and apply my data science knowledge to real problems, preparing me for my future career.
During the spring semester of my junior year, I took a graduate level course: INFO-H 510, Statistics for Data Science. This class covered topics such as probability, sampling, hypothesis testing, and generalized linear models. Each week, after studying the theoretical concepts of statistics, we worked on data dives, applying the concepts to gain insights from a dataset we chose at the beginning of the semester. This work culminated in a final project where we came up with a specific target audience and problem, as if we were working on a project for a real company.
My dataset focused on video game sales, so in my final project, I analyzed the data to help a fictional indie game studio with new investors that needed to decide on the genre to pick for their next game. I used R to conduct my analysis and explained my results in detail on a Quarto Pub site and slideshow. By applying my statistical knowledge to a realistic problem and explaining my work in a business context both helped me understand the concepts better and strengthen my communication skills. Now, as a teaching assistant for the course, I am helping new students become better statisticians and data scientists.
Explore my game sales analysis site below or at this link.
Software Product Operations Internship
The summer after my sophomore year, I was an intern at Allied Solutions in the Software Product Operations department. This small team focused on improving the efficiency of the different software help teams throughout the company by identifying where processes stall, interviewing team members to gain a better understanding of specific issues, and developing pipelines to simplify tasks. My work focused on the first of these tasks: identifying process issues through data visualization.
I met with stakeholders from Product Services and Operations to determine what visualizations were needed, collaborated closely with the Jira product owner to set up access to the necessary data, and worked on building out an interactive dashboard with key metrics on number of tickets received and ticket response time in Power BI.
I encountered challenges along the way, particularly with the time to first response. Jira provides response time data in SLA time, meaning rather than giving the actual number of hours, it reports the number of hours relative to the response time promised by that particular team. To work around this, I had to build out a complex calculation in DAX to count the number of business hours between the start and response times for each ticket.
After creating the key visuals and cards, I implemented filters for priority, team, and assignee, and emulated a filter for project by adding multiple pages. Finally, I worked on documenting the data processing and giving the dashboard a clear, clean design, making it easy to use and understand for individuals across technology and business departments.
To the right, you can view screenshots of the dashboard with different filters applied, filled in with randomized example data.
Screenshots of data dashboard, first with no filters, second filtered to 'Project C' tasks, third filtered to 'High' priority tasks.
MISO Xtern Data Science Challenge
As part of TechPoint's inaugural Xtern Challenge, I was placed on a team with four other Indiana college students and given a task to gather new insights for the company MISO. The goal was to find a way to utilize sentiment analysis to predict changes in energy policy and allow MISO to adapt their strategy accordingly. Being unfamiliar with the energy industry, I used the provided opportunities to speak with mentors from MISO and learn more. One aspect of the company's operations that had caught my interest was that they almost entirely worked inside the United States, but also were involved in the Canadian providence of Manitoba.
After talking with the mentors, I came back to my group with a specific question to explore more deeply: when looking at the sentiment of news articles, which region provides the best indicator of upcoming international energy policy changes? This is an important question to MISO, as they generally import energy from the less populous Manitoba into Minnesota, so shifts in energy policy could impact costs. I planned out a pipeline for scraping, analyzing, and visualizing the data, and collaborated with my team members to divide the work up further.
We chose three specific energy policy events and analyzed the sentiment of articles from national USA and Canadian news as well as local Manitoba and Minnesota news to see if our hypothesis that there would be more positive sentiment before policy changes that increased trade and more negative sentiment before changes that decreased it would hold true. I built out an interactive dashboard in Power BI, displayed below. It turned out that, with the little data we had, no consistent indicators were found. However, our teamwork and potential of our pipeline to be improved with MISO's resources secured our team third place in the challenge.
Screenshots of two dashboard pages, the first highlighting sentiment by region for a specific energy event, the second highlighting key words, articles, and sentiment confidence across multiple regions for the same event.
Reflection
These experiences have been extremely beneficial in the development of my data analysis skills by allowing me to connect concepts with practical problems. I have learned how to handle challenges in data processing, explain the meaning and impact behind my work clearly, and communicate with others to better understand business problems. I feel not only prepared to tackle data analysis and visualization problems in my future career, but also to work with others across technical and business backgrounds in a professional environment.