Daniel Zvinca (@danz_68) profile picture
Joined September 2017
Looking into the evolution of the relationship between two variables. Using the normalized covariance parts to encode the systematic relationship (how the change of one is reflected in the change of the other). Using different smoothing degrees for interpretation.
2
203
Triggered by Carlos Barboza LinkedIn post, linkedin.com/feed/update/urn… this is how a density dots plot design looks on the very same data (yearly cycle perspective).
1
3
309
For skeptics, the accuracy stats: ^max, ~avg, %overlap for Wilkinson vs. KDE
1
1
139
KDE encoding wins by a serious margin to Wilkinson, max error placement is a half and average error is a quarter. The 5% overlapping dots area trade-off goes almost unnoticeable.
1
106
A crystal clear takeaway combining two great dataviz techniques: reading the signal (instead of the noise) and using area for highlighting two line charts difference.
NEW: an ideological divide is emerging between young men and women in many countries around the world. I think this one of the most important social trends unfolding today, and provides the answer to several puzzles.
3
153
Having a bit of fun. As mentioned yesterday, RELATIVE gaps are not properly encoded by a dumbbell chart, which basically encodes the ABSOLUTE gaps. BTW, a bar chart to visually compare relative gaps (comparing ratio of ratios)? A dot chart is a more appropriate option.
1
1
1
4
1,444
In the article there are mentioned the expectation gaps as RELATIVE value (eg. 2.2x for US, also displayed in the table), while the gray bars are used to graphically encode the ABSOLUTE values. Encoding the two distributions next to a sorted by gap table might be easier to read.
How quickly the reader can find the story in dataviz is important: junkcharts.typepad.com/junk_…
2
192
Daniel Zvinca retweeted
Mathematics, logic, beautiful universe. What a 7-way Venn Diagram looks like. By Santiago Ortiz, @moebio, moebio.com, Used with permission.
1
33
120
12,482
Triggered by the existing circle shaped flags, I couldn't resist to try a colorful density dot plot. The arrangement follows the distribution shape. If we admit a slight error displacement, we can reduce overlapping (2nd image). #makeovermonday
1
4
677
A graphical alternative which emphasizes the gaps and avoids overlapping.
Replying to @jburnmurdoch
In Germany and the UK, immigrants and natives are roughly equally likely to be in poverty or in the bottom tenth of earners. In France, just 11% of natives are in poverty, compared to 28% of immigrants, rising to a stunning 40% among the most recent arrivals.
1
7
2,081
Disagree. When data follows a predictable pattern with a trend, a dense cyclic spaghetti chart will hide, in fact, a lot of useful information. A properly smoothed chart will show trend, seasonal pattern variation, exceptions and many more.
This Post is from an account that no longer exists. Learn more
1
1
1
4
1,612
Subtracting the smoothed trend values (light blue curve) see below what patterns are left to analyze. Is the red part that unusual?
1
2
305
Prime Numbers in Polar Coordinates 5,000 (left) Prime Numbers and 50,000 (right) Prime Numbers plotted in polar coordinates.
113
1,006
75
8,116
1,150,098
A makeover of a Marimekko chart for a similar message. A layout with comfortable read/scan and comparison of the most relevant encoded variable: the percentage. One might argue that lengths are easier to compare. True, yet a Marimekko layout hardly gives comparable sizes.
2
490
Daniel Zvinca retweeted
This is a pretty intense (in a good way) treatise on color theory from @NateBaldwinArt. | An interactive guide to color & contrast colorandcontrast.com/#/
5
1
11
4,062
"a little learning is a dangerous thing" Alexander Pope - An Essay on Criticism, 1709
2
1
233