It’s been a while since I’ve had the leisure to play with new open-source tools for data journalism. It’s been even longer since I’ve written a tutorial. Today I wanted to explore a fairly WYSIWYG web app Plotly. Although most of the features on face value appear to be not much more than what you can do in Google Visualization Playground or an open-source library like Highcharts.js – basic chart types on-the-fly, hover interactions – what particularly stands out about Plotly is its ability to perform many of the statistical analyses to your dataset before you go through the process of visualizing it.
Just to give you an idea of the full range of more than 45 different standard chart types, I put together this slick heat map displaying the average unique visitors I’ve received on my blog for the past month during certain times and days of the week. Take it for a whirl:
For example, you can calculate percent change from a set of chronological raw numbers to display accurate trend data by simply choosing “Data Analysis>Percent Change” in the Grid View. I tried this out with some historical data scraped by the folks at enigma.io on the percentage of the Savannah-area workforce in the hospitality industry, even adding in a fit or “trend” line to verify the upward tick: