Makeover Monday this week was about the US Influenza Surveillance Report which tracks the number of patients visiting participating healthcare providers with Influenza-like Illness (ILI) on a weekly basis. The original visualisation was a line graph showing the percentage of ILI patients for 2017-2018 and a selection of previous seasons (a year beginning in October). The original line graph was informative enough but quite hectic with so many coloured lines, my aim on my viz was to utilise Tableau’s filter function to make it so that the viewer can select just one season at a time. Adding to that visualisation there was data regarding the number of ILI patients in different age groups and showed them as a percentage of the whole ILI patient cohort for that season which shows that influenza manifests mostly in those between 5 to 24 years of age and is scarce in older people.
I decided to use Tableau Prep to clean the data, the original data did not use full dates and had a ‘Year’ field and a ‘Week’ field, in prep I used the DATEADD function to add the weeks to the year date so that I had dd/mm/yyyy dates rather than just 01/01/yyyy. Initially I made a small mistake where none of my years would start at 01/01/yyyy since I added the week number given in the field hence week one would be week two and my year ended at week 53 for every year. To solve this, I subtracted one from the week number which sorted that issue. The next issue that occurred, as pointed out by Andy Kriebel (@VizWizBI) on Twitter, was that my drop down was to filter by season, but my year started in January, hence the weeks were not in the correct order. To remedy this, I set the fiscal year start to October (week 40), which is when the season starts in the data – this was easily solved in Tableau Desktop).
Plotting the bar showing the percentage of patients in each age group was a bit difficult since the data had six age brackets, some of which overlapped hence sometimes the bars had null fields which would not display. I used calculated fields to set all nulls as zero, then added the age groups that overlapped to form one larger age bracket. This still left some seasons where the total did not equate to 100% so taking the remainder of ILI patients that were not in an age group I formed a group where the age was not specified so that the bar spanned 100%.
In the Viz Review, Andy and Eva (@TriMyData) gave some feedback regarding some changes that I could do to make the viz easier to read, like having the age group labels within the bar so that it is easier to see which section of the bar is for which age group, and for consistency just keep the line graph one colour rather than changing the colour every time a season is selected.
This week I also had a two-day SQL (Structured Query Language) Advanced course, I attended the introductory course back in week 3 (more details here). The course started with a quick run-through of some of the basics from the introductory course which was useful but, thankfully a lot of the commands and syntax (or lack of) were easy to remember. This course was different to the introductory one, mostly in the structure of the course. This one was much less structured which could be due to the instructor’s personal preference or due to the attendees and what we would be required to know for work. The course had lots of little practice activities again which were a lot more open which were useful; the instructor gave a scenario and what he would like as an output and we had to write up the SQL to get there, this was very useful since it is a realistic situation to know what outcome you require but it is getting there that is difficult. Again, the course provided a book with notes and a PDF with even more content which will be great when I need to recap any details or functions of SQL.
I began studying for the Tableau Qualified Associate exam this week, I started by going over the exam guide provided by Tableau which has a list of the skills they assess. I tried to make condensed but thorough notes on each point in the list of skills with brief instructions on how to perform the skill (if necessary). For me, this was a very useful task since there were some skills that I was unfamiliar with (mostly in the Data Connections section) which I had to search for more information on. This brings me onto Tableau Help as a resource, I used this a lot for the Chart Type skills where I knew how to perform the skill but there are best practice ways to do it (especially when with a client) so that is always outlined in the help section.
For the last part of the week, I did the sample questions in the exam guide to gauge where I am in terms of knowledge for the exam; this will give me sufficient time to go over any parts of the exam I am less able with before I do more practice tests. For the questions in the exam guide I managed to do quite well however the guide does not provide many questions. I did find a source online called Learning Tableau where there are some questions prepared in the style of the exam, this was very useful though I did find the questions quite difficult, particularly the Level-of-Detail questions where the phrasing of the question was not always easy to understand. I will continue to do more practices next week in the hope that I can answer the questions well in a more restricted time frame.
If you are interested in learning more about Tableau or having a free demo,