Powerplay Powerplay data analysis using Inara.cz data

CMDR kekosummer

C
Thanks to a recent update of player journals containing powerplay standings, Inara got an update with detailed information on the system's reinforcement and undermining points.

My addition is in the form of a Kaggle notebook that scrapes data from Inara and runs a simple analysis.
https://www.kaggle.com/code/iananic...lay-data-from-inara?scriptVersionId=240371715 (link to the version created today)

It is scheduled to run weekly every Monday.
You can download it as Jupyter Notebook, copy it on the Kaggle itself, export it to Google Colab...

All plots are interactive, powered by Plotly, and can be downloaded in 'output' tab either separately or in a single zip.
One of the things you can do is change the selection of data shown. For example, if the plot has bars colored by power, clicking on power in the legend will toggle it's display, double-clicking will select only that power.

Below are just some example plots: notebook contains many more.

Overall standings
1747550435911.png


Aggregated points summary
Using progress and power state it's possible to compute total number of reinforcement points held by any given system.
1747550399451.png


On average, at what progress power keep their exploited, fortified and stronghold systems?
1747550359833.png


Reinforcement vs undermining
Compare power's own reinforcement with undermining it experiences (not the one it causes, since that is unknown).
1747550245200.png


Data freshness
None of this would make sense if the data is too old. So there's a section dedicated to exposing old systems, and seeing how age of their data is distributed.
1747550902035.png


P.S. It would be nice to see at least basic plots like these featured on Inara itself.
P.P.S. author holds no responsibility for any offensive actions taken as a consequence of making conclusions based on visualizations presented in this work.
 
Last edited:

Ian Doncaster

Ian Doncaster
You might find it more convenient, rather than scraping Inara, to use an EDDN feed and get the data directly from the FSDJump/Location events in that. (This will also let you get live updates as they happen)

CMDR kekosummer said:
author holds no responsibility for any offensive actions taken as a consequence of making conclusions based on visualizations presented in this work.
Click to expand...
Does it look like anyone is interested in taking offensive actions here?
 

CMDR kekosummer

C
Ian Doncaster said:
You might find it more convenient, rather than scraping Inara, to use an EDDN feed and get the data directly from the FSDJump/Location events in that. (This will also let you get live updates as they happen)
Click to expand...
I am taking an easy route: freshly served aggregated data is nice and simple. Parsing Inara was easy. This is the convenient way. This was done in under 1 day.
Can't imagine the headache needed to set up something that will get events from EDDN and aggregate results, update records, and summarize...
 

Rebel Yell

Rebel Yell
It's cool to have some comparisons indeed... on the other side: EDDN/Inara data are al subject to updates, so these can be outdated.

We need also a better UI in game, where we can sort out the state of all systems without havin to go on Inara or being subject to updates... like it was in PP1 (reinforcements/alerts for undermining, current conflicts/contested + acquisitions etc).
 

CMDR kekosummer

C
Rebel Yell said:
It's cool to have some comparisons indeed... on the other side: EDDN/Inara data are al subject to updates, so these can be outdated.

We need also a better UI in game, where we can sort out the state of all systems without havin to go on Inara or being subject to updates... like it was in PP1 (reinforcements/alerts for undermining, current conflicts/contested + acquisitions etc).
Click to expand...
1) That's why notebook is scheduled to update weekly.
2) ingame UX is... a source of frustration. To some degree I love that community is the one creating tools to understand the game, but at some point I understood that it shouldn't be just that much neccessary.
 

Ian Doncaster

Ian Doncaster
CMDR kekosummer said:
Can't imagine the headache needed to set up something that will get events from EDDN and aggregate results, update records, and summarize...
Click to expand...
Not as difficult as you might think, if you're already storing the data you scrape from Inara somewhere. That said, I already have one of those for my own Powerplay tools, so if you want a CSV/JSON/etc. export of that which you can download a bit more often, then let me know what fields you need.

Rebel Yell said:
We need also a better UI in game, where we can sort out the state of all systems
Click to expand...
Now that they've added the detailed journal updates they might as well do this too.

I did miss the original set up where you could sneakily undermine a system (without sniping) gradually for a couple of weeks because no-one was paying attention to it.
 

CMDR kekosummer

C
Ian Doncaster said:
Not as difficult as you might think, if you're already storing the data you scrape from Inara somewhere. That said, I already have one of those for my own Powerplay tools, so if you want a CSV/JSON/etc. export of that which you can download a bit more often, then let me know what fields you need.
Click to expand...
The neat thing about Kaggle is that I don't store anythin really.
Not an ad of Kaggle, but I like that ability to operate effectively serverless.
Every week notebook runs, makes a few requests to Inara, parses, processes, saves plots, and it's done.
Same with notebooks on stellar data and station. No need to maintain server or storage.
 

CMDR kekosummer

C
Ran update during this week's server maintenance.

Elite Dangerous: Powerplay data from Inara

Explore and run machine learning code with Kaggle Notebooks | Using data from No attached data sources
www.kaggle.com www.kaggle.com

Also added histogram for number of systems per reinforcement progress per power state per power:
1748504952782.png

When all power together it's a bit of a clutter, yes. But when working with interactive plot, you can select power you want to see, and then it becomes tool to evaluate where to put more reinforcement (or undermining...). Logarithmic scale is dictated by significant difference in number for exploited and stronghold systems.
 
You must log in or register to reply here.
Back
Top Bottom