Remake: graficele

Portalul a publicat recent un articol intitulat „Ce arată toate sondajele de opinie pe alegerile prezidențiale?”, interesant de altfel, însă din păcate graficele, esența articolului, sunt pline de „chartjunk” și spații goale (vezi mai jos), având nevoie îmbunătățiri semnificative.Picture2Eror - Copy

Principala schimbare a fost eliminarea simbolurilor aferente diverselor institute de sondare. Ar părea relevante dacă s-ar discuta în articol despre diferențe remarcabile între institutul X și Y, dar așa, lăsate de capul lor, aceste simboluri mai mult încurcă decât ajută. Mai ales că două dintre simboluri (CSCI și ARP) sunt imposibil de diferențiat.

O altă modificare a fost ștergerea sintagmei „Procentajele arată rezultatele sondajelor de la diverse institute de sondare a opiniilor” care pe lângă că era repetitivă („sondajelor … de sondare”) și cam în plus (scopul procentajelor e ușor de intuit), mai era și scrisă la fel de mare ca titlul graficului, înecându-l practic. Titlul trebuie să iasă în evidență, că de aia e titlu.

A treia modificare a fost eliminarea legendei separate, și plasarea etichetelor cu numele candidaților în dreptul liniilor din grafic, pentru o mai rapidă lizibilitate. Și tot în acest scop, am modificat fundalul, din alb în negru, și am restrâns axele pe intervalele pe care există date. (De exemplu, în graficul de mai sus, între „septembrie” și „2 noiembrie” nu există date).

Iată rezultatul:

Redesign 0 Redesign 1 Redesign 2 Redesign 3 Redesign 4 Redesign 5 Redesign 6Grafice realizate în Inkscape

3 thoughts on “Remake: graficele”

  1. Hello and thanks very much for the comments. Alas, I can’t read or write in Romanian but I see what you mean and I like several of your ideas, like cutting back on legends and text, placing the party labels next to their data line, and the introduction and placement of the “in sondaje %” axis label. I am not convinced by the dropping of the info on polling companies (we thought that was an important part of the added value that we create here, i.e. you can easily see in our charts that trends and differences cannot be explained by “house effects”, i.e. differences in results between different companies. You are right that a black background may look better on a computer screen and that we could of course choose a more appealing color scheme. But yet, I would disagree with your redo of the colors because people associate political meaning with colors and candidates make their color choices. Unfortunately, they do not choose colors that would look good together on a chart, in fact most Romanian parties and presidential candidates went for some shades of blue while others use multicolor identifiers, which made our job a bit difficult, but we nevertheless sought to remain as faithful to the color choices of the parties/candidates as possible, and that also means that we cannot pick a black background. So yes, aesthetics are important, and you do well by cutting back on the text that I crowded in the chart at the expense of the data area (next time I will think of that), but there were substantive reasons for some of the choices that you disliked in our originals.


    1. Thanks for the feedback.

      Regarding party colors, I tried to keep as faithfull to party colors as I could, while trying to keep things easily legible. There are only 3 parties that have their color changed: FC and PRM, which are somewhat minor, and ACL, which although blue in this campaign I chose to represent as yellow because a) it stands out stronger in the charts and b) it’s the color of PNL, which is the candidate’s (Iohannis’s) party, and therefore not entirely foreign.

      Now regarding the polling companies, I agree that it’s a big part of the info that I removed, but I actually couldn’t easily see any trends, so without accompanying text to highlight them, I decided to remove them.

      Maybe some further analysis, highlighting bias a bit more, would be interesting. Maybe seeing how it correlates with political links the polling companies have. Maybe something like this:


  2. Thanks again. I will check the color issue more closely when I get a bit of time and I already implemented in the next version some of your ideas. Re the polling companies, I think our main point is that there really are not enough polls published in Romania that one could make anything of whatever – somewhat chaotic – differences can be seen between the various polls. I guess if one finds a small space for the legend for the institutes in the chart, the story is worth about as much as one can see from it in our charts. But will try to experiment with it a bit.


Leave a Reply

Fill in your details below or click an icon to log in: Logo

You are commenting using your account. Log Out / Change )

Twitter picture

You are commenting using your Twitter account. Log Out / Change )

Facebook photo

You are commenting using your Facebook account. Log Out / Change )

Google+ photo

You are commenting using your Google+ account. Log Out / Change )

Connecting to %s