These are chat archives for FreeCodeCamp/DataScience

7th
May 2016
Daniel
@profoundhub
May 07 2016 04:49
@bradd123 i looked into it but decided not to pursue it. are you doing it right now?
Brahma Reddy Chilakala
@bradd123
May 07 2016 06:55
@profoundhub @jboxman I am planning to do. Is it worth the price?
evaristoc
@evaristoc
May 07 2016 07:48

@krisgesling much better! My suggestions:

  • Can you work a legend for the colours? It is not clear what they mean, I assume it is the number of participants who responded the survey per country.
  • Can you try to find a better way to highlight country's borders? White lines are ok for countries in dark green, but for countries in light green the contrast is not enough to distinguish border lines.
  • The information in the tooltip: hmmm... the purpose is fine, but a pie without values in it is not informing... probably the font-size is also too big? The tooltip occupies the whole screen, I would feel it better if smaller... and instead of mentioning the whole gender definition for no-male / no-female I would just write "Other", so the number of words is less. Be aware that writing the absolute values is ok but they are harder to compare. Usually people make percentage available and the SUBTOTAL counts for back-reference.

Let me know if this helps!

Sam Aiken
@SamAI-Software
May 07 2016 13:49

@krisgesling that looks great! Good job! :+1:
Totally agree with @evaristoc about legend, borders and gender definition. Just wanna add a bit:
There is no need for unnecessary text in tooltip, the information should be easy to consume, so we can drop words like "respondent" and sort data descending or just in common order male -> female -> other to make it easy:

- 65% male
- 33% female
- 2% other

Then small black legend box will become smaller, so there will be more space for a graph (pie chart).
Also the total amount of respondents for a country should be more noticeable. We can try to duplicate it in the black legend box. Some like this

- 65% male
- 33% female
- 2% other
-------------
 6258 total

Last approach can look like this:

- 65% male (4386)
- 33% female (1752)
- 2% other (120)
-------------
 6258 total
Sam Aiken
@SamAI-Software
May 07 2016 14:32
@krisgesling check this out. Looks simple and easy to use
pieChartExample1.jpg
Kris Gesling
@krisgesling
May 07 2016 15:14
@evaristoc @SamAI-Software yeah agree mostly the less text the better, especially with the percentages over straight numbers and "n=" probably means absolutely nothing to the majority of people who'd look at this. However I think it's still important to list those who selected a non-male and non-female gender separately to those who didn't choose a gender as they're very different things. I'm not a fan of the 'other' label either. I know it's hard to group a range of different people succinctly without that or an acronym but not impossible to make it work.
Sam Aiken
@SamAI-Software
May 07 2016 15:16
Oh this future :smile:
Daniel
@profoundhub
May 07 2016 16:14
@bradd123 it really depends on what you want to do after you get that degree
evaristoc
@evaristoc
May 07 2016 19:59
@krisgesling excellent point about the Others label. I would keep the map you have done but what about other representation? It can be a bit polemic one, but could you try another in which instead of representing the number of respondents, the colour would represent the proportion of minorities against total respondents? It is a pity we didn't happen to ask about race for example...