I thought I'd better post an update about the Lemmy.nz census. I have been working on analysing the data and writing a post, but there's probably still a while before I'll be done. There's quite a lot to get through.

Also, if anyone knows any good tools for generating graphs then let me know! I'm using LibreOffice Calc as an excuse to learn how to use it, and it's graphs aren't that great. I can break out Excel if I have to but to be fair Excel graphs aren't much better.

[–] absGeekNZ 4 points 4 months ago* (last edited 4 months ago) (1 children)

You could do this with python and a couple of libraries. This is just an example, but you could import the data from a DB or use a CSV file.

import matplotlib.pyplot as plt
import numpy as np

# Pie chart data
labels = ['Category A', 'Category B', 'Category C', 'Category D', 'Category E']
sizes = [30, 25, 20, 15, 10]
colors = ['#ff9999','#66b3ff','#99ff99','#ffcc99','#c2c2f0']

# Pie chart
plt.figure(figsize=(8, 8))
plt.pie(sizes, labels=labels, colors=colors, autopct='%1.1f%%', startangle=140)
plt.title('Sample Pie Chart')
plt.axis('equal') # Equal aspect ratio ensures that pie is drawn as a circle.
plt.show()

# Histogram data
data = np.random.normal(0, 1, 1000) # Generate 1000 random data points with a normal distribution

# Histogram
plt.figure(figsize=(10, 6))
plt.hist(data, bins=30, color='#66b3ff', edgecolor='black')
plt.title('Sample Histogram')
plt.xlabel('Value')
plt.ylabel('Frequency')
plt.show()

[–] Dave 2 points 4 months ago (1 children)

Ooh thanks! I'll definitely have a play with this. What's the step before this? Is it as simple as installing python, putting the code in somefile.py, then running it?

[–] absGeekNZ 3 points 4 months ago* (last edited 4 months ago) (1 children)

Yep, if you are running any type of Linux python is already installed.

I always have a path in my python files to allow for direct running rather than calling python first. This only works on Linux.

If you put
#!/usr/bin/env python3
as the very first line, you can make the file executable and it will just run

otherwise you will have to call python first, e.g. python yourFile.py

[–] Dave 1 points 4 months ago

Awesome, thanks! I'll have a play with this over the weekend.

[–] throws_lemy 2 points 4 months ago (1 children)

I'm curious, can we have something like automation in LibreOffice? I mean like in VBA, where I can pull data and export excel charts.

[–] Dave 2 points 4 months ago

LibreOffice Calc supports its own macro language as well as a subset of VBA! I recently converted an Excel macro to LibreOffice and it wasn't too painful.

[–] liv 2 points 4 months ago* (last edited 4 months ago) (1 children)

Fudge, I forgot to complete it! I got halfway through and started having an existential crisis about my identity ha ha and thought I'll come back later and fix this.

[–] Dave 2 points 4 months ago (1 children)

Well in you hop then, while I'm sorting out how to do graphs. I've removed the expiry so you can finish your submission.

Anyone else can hop in too if you haven't done it, as long as you're quick 🙂

https://survey.lemmy.nz/index.php/493338?lang=en

[–] liv 2 points 4 months ago (1 children)

Thanks! Done!

Great, I'm back to questioning my identity and I'd forgotten the other super hard question in there, dessert. 😃

[–] Dave 2 points 4 months ago

Haha I copied that from the Lemmy.ca one, I thought it would be interesting to see any differences!

If anyone else wants to do it and hasn't had a chance, I'll leave it open until tonight or whenever I next get back to doing the data analysis.

[–] deadbeef79000 2 points 4 months ago (1 children)

Just roll with the LibreCalc charts.

It's the data that's important, perfectly rendered gradients on histogram bars is less important.

[–] Dave 1 points 4 months ago (1 children)

Mainly I don't like that the pie charts are hard to tell what is what. If there are 6 or 7 things and you have to tell which is which by colour, that's a bit tricky. Labels on the sections would be better, but I haven't worked out how to do that.

Maybe if I think a pie chart shows it best, I'll also include the bar chart so it's clearer.

[–] deadbeef79000 2 points 4 months ago* (last edited 4 months ago) (1 children)

Use a bar chart. Pie charts are for marketing and pizza.

Also, no one cares if a bar chart sums to more than 100% ;-)

[–] Dave 2 points 4 months ago (1 children)

I dunno. I feel something like this works better as a pie chart.

[–] deadbeef79000 2 points 4 months ago (1 children)

Pie charts tend to work when you have three to four categories, more than that they fall apart.

The nice thing about the bar chart is the axis label which can be the raw value rather than a percentage, having a large (7-8) number of categories is still readable especially if there minority categories.

Also, is all just my opinion, so don't let me stop you using whatever you like!

[–] Dave 2 points 4 months ago (1 children)

I think pie charts are also helpful for when you have 10 categories but two of them make up 90% of the result.

But your bar chart preference is noted, so I'll try to make sure to include them along side pie charts 🙂

[–] deadbeef79000 2 points 4 months ago (1 children)

Poor data visualisation is a pet peeve of mine and I'm disproportionately vigorous when talking about it ;-)

Especially after a few drinks with dinner.

[–] Dave 1 points 4 months ago

Haha one of my pet peves for data visualisation is graphs that don't start at zero. I get sometimes the movement is too small to visualise properly when you start from zero, but you can't probably understand the scale if you don't start at zero.

I bet you'll argue with me on that!