A few of the data sets I could locate, mostly related to music:
December’s challenge was “Christmas Music” for Dear Data / Chain Data.
I confused the instructions of “dear data” where it mentions “one week”. I understood, that it has to be my personal data collected over a period of one week. Thanks to Bridget (@windscogley) for clarifying it and throwing in some ideas to get me started …
Here is a run down of my journey during the month:
- Songs described in terms of chords or music qualities seemed a good starting point, as described here, https://flowingdata.com/2012/06/20/analysis-of-chords-used-in-popular-songs/ and http://www.ethanhein.com/wp/2011/visualizing-music/ but my lack of knowledge kept me from going that route.
- Music Lyrics, leading to text-based analysis was the next idea. Good pointers for the idea can be found at http://research.dbvis.de/text/ and https://graphics.stanford.edu/wikis/cs448b-12-fall. And a few practical tools here: http://www.tapor.ca/. I downloaded lyrics of a few popular songs, such as “Jingle Bells” and “Winter Wonderland”, and using “checktext.org” recorded certain metrics such as: Flesch Reading Ease, Word Count, etc …. My idea was to see how these metrics have varied over time, but could not find release date/year for classics.
- Music API, hearing a lot of buzz about APIs, I tried with Google Trends, Google News Lab, Spotify, Echo Net along with others but was not able to get any interesting insights.However the following seem to have interesting potential, which I was not able to invest in:
- Internet led me to an all-time hit list across US and UK, a report from a reputed research house. My idea was to see how these songs are faring in the past 5 years, using billboard, or other sites maintaining such stats, but did not take this route either …
- An interesting fact, I learnt during the search was that in 2015 digital (streaming) revenues surpassed the physical ones (http://www.statista.com/chart/3852/us-music-industry-revenues/)
- Another dimension could be to look towards predicting “hit” songs, explained by Brendan Marr, which led me to learn how IOT is being used to turn Crowds Into Light Canvases With PixMob, or how Taylor Swift is leading the pack. Going back to the prediction bit, here is the research, which gives you the “probability of being hit”, I was thinking of taking the 2015 hits and comparing the actual to predicted, but may be another time.
- Some good sources to look for relevant information are:
IFPI.org, BPI.co.uk and RIAA.com.
- And a few inspirations include:
- Music Genome Project, not much details, but looks interesting …
- I settled for plotting 2014 stats of music consumption across the various genres. Here it goes.
- The data set from Neilsen’s 2014 Music Report for US:
I calculated the index using min max difference, as follows:
And then created a “Parallel Coordinate Plot”
Followed by “hand drawing” it, I only drew the points, excluding the lines for clutter:
On the lines of dear data (http://www.dear-data.com/) where two courageous women
switched continents get to know each other through the data they draw and send across the pond
Brittany Fong (@BFongData) took the initiative to take it wide and far. In her words:
Contrary to how we normally think about data, dear data isn’t about the numbers being visualized as accurately as possible. I mean obviously we don’t want you to just make stuff up but the expectations for dear data are about stepping out of the box, doing something you’re not used to, and being creative
I have joined the bandwagon, and this is what I could muster to draw by hand … yeah it takes a lot of courage to ditch your favorite visualization tool
Here is the same thing, done by the machine
(I was only able to recreate the third chart by hand, using slightly different color scheme)
In summary, we collected food related data for a week, and then analyzed it. I used the Australian Survey Classification System to classify all food items and then presented it in a stacked bar chart
Here is the summary stats for the chart:
and the same summary stats, on the back of the chart