Just a little glimpse of a data story I’ll tell ’Bout a North Country singer that you all know well


Screenshot of the Tableau Dashboard. Available [here] and at the end of this post.

A dream of Bob Dylan
When I was a kid, my parents wanted to ensure that I’d do as well as I possibly could in exams and get a place in University. To this end they hired a personal tutor to give me additional tuition in a variety of subjects that I kinda sucked at.* That’s why they hired JS. The routine was always the same … JS would come to my house and I’d attempt to feign interest and proficiency in my coursework for an hour or so, a couple of times a week. One evening I was listening to the radio when he came over and his first words were ‘What’s that on the radio?’ … well, no, actually … he didn’t say that at all … his language was peppered with obscenities and, when I said it was just something from the charts, he only got more agitated. Lessons were abandoned for the evening as I was treated to a tirade on the poor quality of what made the charts (it was the mid 80s … he wasn’t wrong). The next week he brought around a selection of ten albums on vinyl – exclusively 60s and 70s stuff – and I was instructed to listen to five by the following week. The next week he would arrive with a further five to bring me back up to ten albums and I’d choose again. After a while, I began to develop an interest in Young, Waits, and especially Springsteen. After I ran through all the official Springsteen recordings JS started plying me with bootlegs and the vinyl stack was exchanged for a bank of audio cassettes. At some stage he included a single, solitary Dylan bootleg. I’d heard about Dylan … he couldn’t sing. I knew that much. Why bother listening to a guy who couldn’t sing? I listened to more and more bootlegs. Some I copied and some I gave back … and the Dylan tape stayed there on the shelf. I just wasn’t interested.

I think I only played the tape so I could say that I’d listened to it and could give it back to JS. I remember that it was a sunny afternoon and the house was quiet and still. I don’t really know what I expected to hear … but it wasn’t that! What I’d just hit ‘play’ on was a taped version of the ‘Passed Over And Rolling Thunder’ album and the first track was the performance of Maggie’s Farm from the Newport Folk Festival on July 25 1965. This was Dylan’s first ever electric public performance and the crowd were less than impressed. From a shouted ‘let’s go!’ Dylan, with the screaming guitar accompaniment of Mike Bloomfield, roared through the track and on into Like A Rolling Stone and It Takes A Lot To Laugh, It Takes A Train To Cry. It’s one of the iconic moments of 1960s popular music and much has been said and written about the events that evening and their wider cultural significance. Back then, I had no idea of any of this. I was simply blown away by what I had heard. As soon as I’d played through the tape, I flipped it over and started through again. Right there – standing in front of my twin-tape deck stereo in my bedroom in rural west of Ireland – everything changed in that one moment. After that, Springsteen, Young, Waits, and all the rest didn’t quite mean so much. I listened to every album I could get my hands on – and in whatever order I could get them … which is why I heard Knocked Out Loaded before I listened to Blonde on Blonde. I remember reading Anthony Scaduto's ‘Bob Dylan: A Biography’ where he picked through the meanings, origins, and influences of tracks from Highway 61 before I’d heard them myself. I resolved to get a copy of the album as soon as I could and hear how these lyrics sounded as sung, not just as words on the page.

It all went from there. With JS’s help I somehow managed to scrape my way into university, reading archaeology and English literature, but the Dylan obsession never waned. While there I met JG - one of the greatest and most inspiring teachers I’ve ever had the good fortune to encounter. In her all too brief time at that university, she organised a competition for English literature students called the ‘Reading Prize’. The idea was to work on a topic of your own choosing over the summer between second and third years, and deliver it as a lecture in September to a board of academics from the department. I spent the summer researching around the topic of Dylan’s use of feminine archetypes from his artistic muse to his conception of the Christian church – a topic which still excites my interest to this day. The lecture was well received and I was proclaimed the winner of the competition. I developed the theme further in my final year dissertation – the dreaded ‘Extended Essay’. With that, I managed to scrape my way out of university, but my academic interest was devoted to archaeology by that point. For the best part of two decades I attempted to insert Dylan references into the titles of papers, reports, and just about any other piece of professional writing or lecturing I could manage. Eventually, many of my editors got wise and excised my thanks to various B Dylans and R A Zimmermans from my acknowledgments … but a few did get through!

The Dataset
When I changed career in 2011, it was inevitable that I’d find some way to incorporate Dylan into my work life. Somewhere along the way I was introduced to Tableau as a tool for producing interactive visual analyses and I ran with it. My initial projects were small-scale and fairly simple, but I needed datasets to work on to both improve my abilities and showcase my skills. I found that – unsurprisingly – I did my best work and learnt most when I was using a dataset I cared personally about. In retrospect, it was inevitable that I’d come back to Dylan! Using the data available on the bobdylan.com website, I initially compiled an excel spreadsheet of all known Dylan performances. The data was recorded at the level of one line per song and the details included, Performance Date (broken out by Month, Day, & Year), City, Country, Venue, and Song Title. I’ve expanded the dataset and checked various parts of it against the data held by Olof Bjorner’s bjorner.com and Bill Pagel’s boblinks.com, along with comparing the setlists against my own collection of bootleg recordings. A second dataset listed every song on every studio album. The details included the Album, the Year of Release, and the Title. I made the decision here not to include songs like I shall Be Released which did not get an official release until many years after it was originally recorded. On the other hand, I have included the 1975 release of the Basement Tapes as an original album even though the songs were all recorded several years earlier. I’ve also decided that cover versions become Dylan songs when they appear on official Dylan albums. For example, until 2017 (when Dylan released Triplicate) songs such as I Could Have Told You (written by Jimmy Oliver and most famously performed by Frank Sinatra) were regarded as cover versions, but with the appearance of the studio album are now ‘cannon’. I make no apology for how I’ve decided to order the data, but you should be aware of my biases and understand that there are other ways of constructing the same dataset. Whatever you think of my choices of selection, the core point I want to make here is that the following dashboards only make use of the songs from officially released albums. The significant corpus of original Dylan songs not committed to an official release, cover versions, and traditional songs are all recorded in the dataset, but are not used here … maybe one day I’ll get around to doing something more with the data. The final dataset I constructed was a list of all the cities where Dylan has performed, along with their Latitude, Longitude, Country, and Continent. I’ve taken to appending the US state to the city name to avoid confusion with Old World locations (e.g. Birmingham, AL vs Birmingham in the UK). I’ve done the same for Canadian Provinces and a number of Australian States.

The dashboard
Every time I share what I’ve come to call ‘The Big Dylan Dashboard’ someone invariably asks ‘How do you use it?’ … I’ve ended up writing and rewriting lengthy posts in Facebook comments, so I thought it was time to present something ‘official’ to act as a User’s Guide. The first thing I’d note is that the current version is designed only for viewing on a desktop machine. I’ve viewed it on a tablet and it looks rubbish and it’s pretty well unusable on a mobile phone. I may yet go back and attempt to rectify this, but for now you’ll get the best results on a regular desktop computer.

The basic layout shows four graphs – one map and three bar charts – with a ribbon of album covers along the top and a set of filters along the right-hand edge.

Through hostile cities and unfriendly towns …
The map (top left) is probably the most visually distinct of the graphs and the easiest to read. It is simply a case of one dot per city where Dylan has performed. The colours are by continent and as I reckoned that was pretty obvious, I’ve not included a separate colour key. Where space allows, Tableau will add the city name for clarity. The hover-over ‘tool tip’ gives the city name along with the Venue and Performance Date. However, the last two are usually represented as an asterisk when looking at all the data together as there may be multiple venues and performance dates for individual cities. I’ve left this in as it becomes more useful when the user narrows down their search to smaller chronological spans.

Selection tools on the Map

The theme of this dashboard is interactivity and it’s something Tableau does rather well. It’s a point I’ll be making repeatedly throughout this post, but it does bear repeating – if you select something on one graph to refine your search in any way, all of the other graphs change to reflect that change. Thus, if you randomly click on one city (say, Reykjavik, because it’s off on its own and easy to spot) the map goes dark for all except that dot. And the other graphs change accordingly, just filtering down to the two nights played in the city (1990 and 2008), the albums Dylan played from and the songs he sang. Try it for yourself … I’ll wait!

If you want to select more than one city left-click anywhere on the map and you can draw a rectangle to take in as many or as few as you please. The rectangle is the default, but if you place the cursor anywhere on the map a set of controls pops up in the top-left corner. The bottom-most of these is a caret/arrow/’play’ button (however you prefer to describe it). Clicking on this will allow you to Pan, Zoom, and set the selection tool to either a circular or irregular area – and back to rectangular. Not surprisingly, the ‘+’ symbol zooms the map in, while the ‘-’ symbol zooms it back out, and the little house icon resets everything to the original top-level view. The magnifying glass allows you type in place names and have the map zoom and centre there. The caveat here is that this is searching the OpenStreetMap data, not the Dylan dataset, so does not filter the other graphs. For example, you can search for Newry and the map will zoom and centre itself there, but there will be no dot to show that Dylan ever played there and the other graphs won’t change. The other point I’d make here is that you can see – once you zoom in close enough – that I’m mapping to the centre of the city (as supplied by Wikipedia), not the actual venue … even I don’t have the time or the patience for that level devotion/crazy. Finally, I’d point out that although some schools of thought would place Israel and at least some of Turkey in Europe, I’ve placed both countries in Asia. I’m not making a political statement in either case.

Night after Night ... average age of performed songs

Night After Night
While the map may have been the easiest to understand, the Night After Night graph may be its opposite. At the top level – looking at all the data – it shows the average age of the setlist on the night it was performed. For example, if Dylan played five tracks from Blonde on Blonde in 1966 (the year of the album’s release) they’d be averaged to 1966 and then have the year of performance (also 1966) taken from the answer, leaving 0. However, if he played the same five tracks from Blonde on Blonde in 2006 they still average a release date of 1966, but when the year of performance is taken away the result is -40, indicating that (on average) the songs are 40 years old. In this way, the same set performed on different nights – years apart – will give different results. My original point in creating this graph was to examine the charges of Dylan’s eternally unchanging setlists. The idea is quite simple – where the bars are the same length over long periods it indicates that the average age (if not the exact setlist) remains constant. When brought down to individual albums or songs, the graph can show the gradual ageing of a piece. Occasionally, when the bar appears above the line as a positive figure it shows when pieces were performed in concert before being officially released. For example, the November 1979 run of concerts in the Fox Warfield and other US locations all have an average of around 0.438 years as the setlist was heavily packed with songs from Saved, not released until the following year. Like the map, this graph is coloured by continent.

Top 20 Albums 
Top 20 Albums
This graph is simple enough – it represents the most popular albums by the sum of the songs performed from each. This is, of course, influenced not just by the number of songs on any particular album, but the number of times it is played. While Highway 61 Revisited is the most popular album overall, this is in large part due to the frequency of plays of Like A Rolling Stone, Highway 61 Revisited, and Ballad Of A Thin Man … the two performances of From A Buick 6 … less so. Click on individual albums and the rest of the dashboard changes too and Ctrl+click to select multiples. Obviously, if your selection brings back less than 20 albums, that’s all that will appear (I feel I have to add this after being challenged over why were less than 20 albums listed when only performances from the 1960s were selected *sigh*).

Top 20 Songs
Top 20 Songs
Like the Albums, this graph on the bottom-right shows you the most popular songs whether it’s All Along The Watchtower over his entire career or Lay Lady Lay from the Nashville Skyline album.



The Filters
You can do all the selection you like by clicking on the graphs and seeing what happens, but if you want to get fine grained about your search through the dataset, you need a different approach. That’s why I’ve included additional filters. Again – change something on one of these and all the graphs change accordingly. At the coarsest level, you can search by Decade and Continent. If you’re only interested in Dylan’s performances from Europe in the 1980s it’s only a couple of clicks away – just click ‘all’ to clear out the filter and then click to include what you want. Admittedly if you’re looking for his appearances in South America in the 1960s, you’re out of luck, which is why if you select the 1960s under the Decade filter, South America isn’t available in the Continent filter.

If you want to get down to individual dates you can, using the callipers slider. Just move either end to change the date selection. You can also click on the beginning or end dates and use the calendar function, or type it in manually – all your choice! While it can be difficult to select just one favourite Dylan song, there can be times you just need to know the four locations and dates where 2 x 2 was played (from 1990s Under The Red Sky) and the Search by Song filter gives you just that! The Search by Country and Search by City filters work in exactly the same way, allowing you to get to the data that’s most meaningful to you.

Below the filters are three small panels, each bearing a number. These are included to give you a quick reference to the amount of data in your particular selection and record the number of Unique Songs, the number of Performance Days, and the Total number of Songs.

The Album Cover Ribbon
The ribbon of album covers along the top edge of the dashboard isn’t just a pretty design (though I do love how it looks). As you make your choices in other parts of the dashboard, this ribbon expands and contracts to reflect the albums included in the selection. You can also click on individual covers to refilter the dataset to that one album (again: Ctrl+click for multiple selections).

It’s also important to know that, no matter what you do, you can’t mess things up too badly! If you ever get in a muddle over what you’re filtered or where, you can just got to the bottom-left margin and hit and of the Undo, Redo, or Reset icons and everything will be restored.

If you like this dashboard and find it of interest, I’d ask you to consider doing two things. First – share this post with other Dylan fans or data fans (perhaps even both) so that they can get a kick out of it too. Second, go check out some of my other Tableau dashboards and see what you think [here].

Finally – You can take as serious and scholarly an approach to this data and this dashboard as you choose, but the best way to make discoveries is to just jump into the dashboard and find your favourite Dylan … you may discover something that interests and intrigues you!

Notes:
The title is taken from the Dylan song Hard Times In New York Town … but, of course, you knew that.

As with any endeavour, there is always the chance – nay, certainty – of error. I am particularly grateful to so many Dylan fans with greater knowledge than I possess who have noticed mistakes of various kinds and kindly pointed me in the right direction. The dataset and the dashboard is so much better for their input. In particular, I would like to acknowledge Eduardo Ricardo for his deep knowledge of all things Dylan and his gentle advice.

At the time of publication, the dataset ran from December 24 1956 to April 17 2017 and I plan to keep it updated and revised for as long as Bob keeps touring and performing.

* Including ending sentences with prepositions.

If there are issues with this embedded version, try the dashboard on my Tableau Public page [here]


Comments

Popular posts from this blog

Renewable Heat Incentive (RHI) non-domestic beneficiaries: an interactive analysis of the data

Notifiable Infectious Diseases Reports (NoIDs) Northern Ireland | Trends & Predictions