Presentation showcasing my exploratory data analysis for the popular streaming platform Twitch.
During my masters program, I took a class for my User Experience specialization in Winter of 2022 that focused on principles regarding interactive information visualizations. In this class, we had a module that discussed exploratory data analysis. Exploratory data analysis (EDA) is a method of analyzation of datasets and producing data visualizations that provide more discovery and insights than traditional statistical hypothesis testing would (confirmatory data analysis). EDA combines Data Science and Data visualization to empower analysists to use models and create charts and graphs to discover unexpected findings in the data and suggest possible explanations for causes of observed features.
For this project we were to conduct an exploratory data analysis of our own and present the process and findings in a presentation. We had the ability to choose any dataset we wanted, and since I always try and incorporate my interests into my work, I chose a data set revolving around the popular gaming streaming platform Twitch. Attached below is my exploratory data analysis presentation. I all of the visualizations created in this presentation were made using Tableau. You can see the data provenance, the data prep and various questions to consider as you work through my analysis and finishes with my final conclusions from this process. Looking back on this project a year later, I find the methodologies behind EDA very interesting. The combination of data science with visualization tactics to assist in the information discovery phase of the data process is an interesting way to find hidden variables that might have gone unnoticed if previous statistical assumptions about the data were made (that of which can occur in confirmatory data analysis). Now in my second year of the program and pursing the business intelligence track, I think EDA would be an interesting tool for BI analysists in addition to data scientists to use given the methodologies of EDA relying on both front-end and back-end data manipulation and visualization tactics. I think my project would have benefitted from different fonts in my graphs and a different color palate as well. However I find my insights from this EDA were informative and conclusive, and I believe EDA is a valid methodology to deriving valuable information from data and Is a valuable tool to have in my BI and UX toolkit.