In this blog, I am going to look back over my final application dashboard and show what I could have done differently.
(UPDATE: After writing this blog, on Friday 16/12/2022, DS37 were tasked with updating one of their application vizzes. I chose to update my Final Viz. This blog now includes what I wanted to change on my Final application previous to my presentation, and what I actually changed. The changes will be in Bold Italics)
To start, I should have prepped and cleaned my data set before even attempting to visualise the data. This could have been done using Alteryx or Tableau Prep. In doing this, it would have saved a lot of time trying to figure out what certain measures were and in turn made my life a whole lot easier! One part that I struggled with was the field 'State Name' as originally it was abbreviated. After playing around with Tableau I managed to add a join in which changed the name to the Full state name, but what took me hours to do could have been an instant join if I had prepped my data! I also included a supplementary data set to show the sales profits of the leading chains, and again if I used a prep tool I would have saved myself a lot of time and effort!!
The next change would be getting rid of the pie chart at the start of my Viz. I included both a pie chart, and a bar chart which the user could switch between. Looking back, this is a bit unnecessary and one would have sufficed. Once the pie chart was removed, I added a header to show which bar was Independent Restaurants and which was Chain. Carl recommended that I then turn the chart to the side in order to make it a bit more appealing.
Another part of my Final Viz that I am not a fan of now is the amount of words I used. At the time, I was very eager to get my points across and to guide the user in finding the analysis. But, I have learned that less is more and that big chunks of text are harsh on the eye. Visually, I think I could have made the font a bit more interesting. I should have added some definition within my Titles, and used these to lead some of the analysis I was providing. I cut down the amounts of words that I used on my Viz and made it a bit less explanatory. I also changed the font so that it was the same across my whole dashboard.
In my Viz, I included a moving reference line to show the 'Chainness Rate' (Percentage of Chain Restaurants per State) and the 'Population' of said State. I think this worked really well and enabled the user to interact with the data. I should have included an average reference line so that the user could see if the selected state was above or below average. I took this idea and added it into my scatterplot. There is now an average line for both Population and Chainness. One thing I think I could do further is to colour the different quadrants so that they change if the selected state is above or below the average for chainness/population.
Looking back, I think I should have changed my final Map, which shows the most popular cuisines of each County and the density of 'Chainness' in each County, into something else. Or, I could have changed it so that the colours blended into each other more, as I don't think it gives an accurate portrayal as to the rates of 'Chainness' in each county of the USA.
After a lot of thought, I decided to take a different approach to this map. I don't think the colour added anything to my dashboard and it was quite harsh on the eyes. In the end, I decided to change the map so that it showed the most popular restaurant type (Chain or Independent) for each county. This map looked a lot better and allowed me to include the most popular cuisine of that county in the tooltip also. Lorna recommended I split the chart into two so that we could see both the Chain and Independent Restaurants side by side!
I added an extra bar chart to show the Top 5 Counties in the USA that have the most varied cuisine type. This was split between Chain and Independent and I believe it could add a lot of interaction to my dashboard. A suggestion I had was to allow the user to select a county on the map and the chart would update accordingly.
The last bit of my Final Dashboard that I would change is the conclusion. As mentioned before, my Viz is pretty wordy and I think my conclusion shows this the most. I could have used analysis more to conclude my points instead of a huge chunk of text. If it was more snappy and engaging then I feel the user would find it more enjoyable to interact with. I decided to stick with my wordy conclusion as It summed up my analysis nicely. I did add a show/hide button to allow for a bit more space on my dashboard.