I follow Storytelling with Data on LinkedIn. A great organization, which has created a vibrant community around best practice principles of telling stories with data. Every month they post a new challenge, and their passionate community always comes up with great solutions. These are fun and I learn a lot from all the fabulous submissions but I haven't yet taken the leap to get involved myself. Until the January challenge.
January's challenge was to 'Visualize your Resume'. I didn't actually take part but I was inspired by those who did and it got me thinking: "How could I do the same using Power BI?" (of course).
So here is what I created.
A Powered UP Resume on Power BI
Click on the report expand icon to get a better view of the report
Click here to see the resume in a separate browser window.
April 2021: Star Wars - a New Resume - I created a new version of this for a specific opportunity where I found out that the hiring manager was a real Star Wars fan. There are a few Easter eggs in this one which hopefully make it fun.
Design Prep
The first step was to grab my current resume and consider how I could rebuild it in Power BI. What elements might be more beneficial to see visually vs what would still benefit the most from being delivered in text?
Always taking the opportunity to put into practice the LUMA techniques around Human Centered Design, I pulled out my LUMA Lunchbox with my pack of reference cards, and decided to look at my resume through the lens of creating a concept poster.
I then quickly sketched what my Powered UP resume might look like in Power BI.
Data Prep
Getting the data was fairly easy. LinkedIn has a facility to request your data for download which gives you a number of csv files. I used these as the principal data sources to build the model. Additionally, this was all publicly available information on my LinkedIn profile which gave me the opportunity to represent what was already online, but in a different way.
The Build
I tried to break the resume in 4 clear sections. My Profile, Work Experience, Timeline & Recommendations by using white space to draw clear delineation. Each section takes up the full screen when the title is at the top. This was by design to allow the resume report to be on one long portrait page, rather than spread across multiple report pages.
Here is a short summary of each section and what I used to build it:
The Beard
I thought I would have a bit of fun with this resume whilst trying to show some of the capabilities of Power BI. I figured that if I could get some level of engagement with readers, then it might make my resume more memorable. Thus I came up with the concept of conducting a survey on the state of my face and give them the opportunity to vote.
I set this up using Microsoft Forms so when the reader clicks on the 'Vote Here' button, it launches a new browser window with the ability to submit their vote as seen below.
Power Automate then pushes the results to Excel, which in turn loads into Power BI. With Excel being in OneDrive for Business, a web connection is used to ensure that data refreshes can be automated.
Now for my first use of a visual. In this case, a donut chart worked well. It will be interesting to keep abreast of what people vote here.
The Map
I had seen a few other visual resumes that listed the places lived and I decided to do something similar. Power BI has a few mapping options, but for simplicity, I used the standard map visual.
Again, thinking about the FUN element, I decided to include a fun fact on the customized tool tip. I created the background image and fade using PowerPoint and then used a multi-row card and table to display the data over the images.
My Work Experience Section
I thought about different ways to work with this part of the resume but the reality is that story here is best told in words, rather than visuals. I did try to make it a little less wordy by using the company logos as well as text boxes to show the key dates and length of time I worked at each.
In an attempt to make the page more engaging, I hid most of the details for each role in a pop-up box. This was accomplished using bookmarks and editing the action of the buttons to show and hide on-page elements.
TOP TIP: To make working with bookmarks easier, ensure that you name all of your charts and shapes with descriptive titles, even if not displaying the title in the report. I would highly recommend doing this as you go rather than waiting until the end.
Grouping makes administration of bookmarks much easier. However, grouping does come with its own frustrations when trying to move objects around on the page. In some instances, when you move or resize an object, it can be easier to first move it out of the group by dragging and dropping it in the selection pane. You can then move it back into the group once your changes have been made.
My Timeline Section
I used a similar bookmark/pop-out concept to drive the timeline feature. Each section has its own icon which I created in PowerPoint and imported as an image onto the Power BI canvas.
In my opinion, Power BI has never had a great ability to chart timelines. Until now, maybe? I used the custom visual 'Timeline by Queryon' and I was pleasantly surprised with the result. It would be great if we could have more granular control over what happens when images or text overlap but this is a pretty good custom visual which I think many will find a good use for.
My Recommendations Section
If you are lucky enough to have Power BI Premium, then you have access to use Azure's Text Analytics feature which is part of Cognitive Services. This makes it extremely easy to run sentiment analysis across blocks of text. Perfect for use on the recommendations that I have been fortunate enough to receive on LinkedIn. First up, I needed to clean up the text that had been exported from LinkedIn. For some reason any use of ' or " resulted in the reproduction of some unwanted text characters in the csv. Take a look at the following example;
Great work with our key clients using your ‘consultancy’ mindset to provide strategic thinking on the use of data and analytics to drive decision making. Has delivered real results, well done.
Luckily issues such as these are a cinch to clean up using the Power Query Editor and a little bit of M. The following M code had been used to remove all unwanted characters leaving the text ready for further analysis.
= Table.AddColumn(#"The name of the last step", "What you want to call the new clean column", each Text.Select([The Column with the text to clean], {"A".."z", "0".."9", " ", "," , "-", ",","."}))Next up was how to display the recommendations. I used the Card Browser to achieve this. I liked the fact that I could include the LinkedIn profile pictures which I thought gave it a little more of a "professional app" feel. The word cloud is a tried-and-tested way to show word frequency. Additionally I added a few cards around word count and sentiment.
Don't Forget...
Once I had completed the layout, I set about tidying it up before publication. A key action here was to create the tab order of all the elements. This started off as a bit of a tedious affair. However, I found a little trick;
TOP TIP: If you have a lot of objects, hide them all first. You then unhide them in the tab order you want. This saves a lot of time and is much faster than using the move up/down arrows to set the tab order.
Final Thoughts
All in all, this was a fun project which took about 3 days to complete (2.5 days on design layout). Extensive use of bookmarks were used to give it an app like experience and hopefully the little quirky features will make it memorable to likely employers. It wasn't possible to Power UP all aspects of the resume but overall I am pretty happy with result.
But the proof will be in the outcome.... what do you think? Will Powering UP my Resume with Power BI help me get a job?
Would like to hear what you think.