We chose Chicago Taxi data that has been published by the City of Chicago. The dataset is pretty large with 105 million records and we chose it because it gives people an opportunity to work with data that contains geographical fields, dates and payments data and has plenty of scope for various types of visualisations.
Yes, there are A LOT of rows in this dataset, but the fields are pretty simple and the measures are limited to distance, duration, fare components and number of trips. Nothing complicated and hopefully something everyone can relate to.
The original viz comes from an article on transportation in Chicago and looks like this:
What I like about it:
- simple line chart with a single colour showing taxi trips and taxi fares over time (by month)
- the y-axis has been truncated to highlight the movement in trips taken over time. This helps to see monthly changes more clearly
- while I’m usually not a fan of gridlines, I think they actually work quite well here to divide up the chart area and guide the viewer to the values
- the chart titles are clear and concise and tell the viewer immediately what they’re looking at
- the data source is clearly stated at the bottom of each chart
What I don’t like:
- the data is monthly but it is really hard to identify the individual months in the chart because there are no labels. The heading of chart 1 and my crude counting attempt makes me assume that the vertical gridlines show the end of each year rather than the middle or start. But I would like to know for sure, so some form of labeling or minor tick marks would help
- does the second chart (taxi fares) start in Jan 2013 as well? I guess so but I can’t be certain.
- I would also appreciate some labels along the line for the peaks and troughs in the data, especially those data points that are referred to in the article
- the axis labels could be shortened with an ‘M’ suffix to save some space
- the article states the data release following the Freedom of Information Act request included data from January 1, 2013, to June 30, 2016. Does that mean all of that data is included in these charts? I can’t figure out the range just by looking at them
What I did:
- this time I wanted to go mobile again but with a landscape layout rather than a long form dashboard
- My design inspiration actually came from the appstore and an app called 303 Taxi, which is a provider of Taxi services in Chicago
- I started my design process on paper because I was actually really pressed for time yesterday and needed to have a really solid draft version before starting to work in Tableau to bring it all together
- given that I had my paper version, creating the actual viz wasn’t that time consuming (formatting is another matter though…). I simply followed my draft and adjusted the content based on what I found interesting in the data. Along the way I updated the colours and formatting to what I wanted to final viz to look like.
- given that I worked with the 105 million record data set, I couldn’t just publish my viz to Tableau Public, because workbooks there are limited to 15 million records of data. So I pointed my finished dashboard to the extract from the Makeover Monday website and changed a couple of charts to ensure there was still an interesting story (e.g. the highlight map on the right of the original viz is limited to a single row with the extract because we only included a single community area for pick-ups in that extract).
- following my adjustments, I published the new workbook to Tableau Server, while the original viz can be seen below:
- If you want to check out the interactive viz on Tableau Public, please click on the image below.
Please note: Unfortunately Tableau Public on mobile browser doesn’t allow for scrolling sideways, so this landscape format has been a complete fail from that perspective. I still like it. I might create a vertical version once I stop being annoyed about it…