Guest blog post by Divya Parmar
After data science, which I discussed in an earlier post, data visualization is one of the most common buzzwords thrown around in the tech and business communities. To demonstrate how one can actually visualize data, I want to use one of the hottest tools in the market right now: Tableau. You can download Tableau Public for free here, and the “Cat vs. Dog” dataset can be found here. Let’s get started.
1. Play around with the data and find what looks interesting.
After opening Tableau Public and importing my Excel file, I looked over my dataset. I was curious to see if there was relationship between the rate of cat ownership and dog ownership. So I put dog ownership on the x-axis and cat ownership on the y-axis; I then added state name as a label. All of this is done through simply dragging and dropping, and below is a snapshot of how intuitive it is.
2. Add some elements as necessary to show your insight.
There are many ways to build on the preliminary step. You can add something like a trend line to demonstrate a statistical relationship (note that there is a p-value with the trend line), which is done through the "Analysis" tab and adds more credibility. You can even give different colors or sizes to different data points, as I have done below using the number of pet households by state to emphasis the larger states.
3. Fix and improve to make usable for export, presentation, or other purpose.
Data visualization is only useful if it is simple and to the point. In the above example, the District of Columbia data point is an outlier that is making the rest of the graph harder to read. You can edit your axis to not show D.C., and can also remove the confidence bands for the trend line to remove unessential information.
After your visualization is ready, put it to use by sharing, embedding, or whatever means works for you. Data visualization is easier than you think, and I encourage you to get started.