Displaying weather data nicely

I have a huge spreadsheet with weather data (~200.000 rows, 18 columns). I’d like to display this data in a nice graphical way. I tried libreoffice (calc and base), but it’s slow and not nice graphically. I also tried labplot (which gave me headaches with imports first), and somehow managed to get some graphs, although the display still isnt that nice (Zoom levels, smoothing, etc takes a long time to get to understand… :wink:

So, any tipps on what to do, what program to use, and how to format it so that the graphs look nice? Or is it too generic a question?

Calc, Calligra, OnlyOffice, WPS, FreeOffice Planmaker (not all FOSS BTW) are options…

I can’t say what’s best, not much experience there… but it might be worth installing and testing them.

There’s other stuff, like Matplotlib, Ggplot2, etc… for ‘data plotting’ tools.

We are talking about several millions of data points. I would learn some basic python and explore some solutions in that direction (matplotlib, seaborn, plotly).

It also depends on whether you want to produce static plots or you need an interactive view.

2 Likes

this is not much data for LabPlot. which kind of problems do you have when trying to achieve nice looking results in LabPlot? If you can provide concrete points, we can definitely help.

Ok. I’ll try. My weather station uploads updates to the internet every ten minutes. I can download those for each month. So I have monthly csv files with data about temperature, humidity, wind speed and direction, rain fall and uvi.

I would like to be able to look at weekly or even monthly figures to compare rainfall and other patterns over the years.

Right now I have labplot set up to use a csv as a life data source, and I concatenate all the monthyl csvs into it. So that works so far. It’s just that I don’t know how to add those ten-minute-spaced values into a weekly or a monthly sum.

Sorry for sounding so complicated, English is not my first language and I don’t really know how to explain this better…

What you’re looking for is called down-sampling of a time series. This is not something that can be directly done in labplot yet and you’ll need to use another tool to resample your data. Here an example with Pandas:


Here, I used the data set from Max Planck Weather Dataset | Kaggle which provides the data for every 10 minutes and I created a new time series that has average values per week. After the extraction of the columns, conversion to arrays, you can in principle plot these arrays labplot and use those arrays as if they were usual columns in the spreadsheet.
I just realized, we’re not parsing the datetime correctly. So, it won’t be possible to plot those arrays directly. We’ll fix this problem. But I hope you get the idea. Just pre-process your data in another tool and import the data into labplot that you want to visualize. Here another tutorial for how to do this in python only - https://machinelearningmastery.com/resample-interpolate-time-series-data-python/.