Developing QuantFM, an app for FM quantitative analysis - help wanted

[gam945]

I have in the past already developed an FM companion for the more data oriented managers in the form of a web-app, FM Data Analytics for FM21, which got quite some engagement from the community at a time were statistics were under-used in the game UI. This app is not available anymore, although I still have the source code, but if it certainly had quite some good features for its context in the game development, it also presented some big cons, for example:

• Deployment: users had to download the source code from my online repository on GitHub, install R and make sure the installation directories matched what the app and the batch script to launch the app expected. This was way too complicated for the average user who'd like to use the app with minimal involvement from his part.
  How QuantFM will correct this: the app will be available as a downloadable software - download an installer, which will install the app and make it usable as is and offline from your computer - or will be hosted online.
• Pre-set analysis: users were able to select from multiple, varied visualizations but in order to get custom visualizations, the source code had to be modified. This maybe was trivial to some users, but definitely not to many others.
  How QuantFM will correct this: pre-set analysis will be offered but users will also be able to build their models in-app using the UI only - for example, change the formula or weights used to calculate a position/role/duty score from statistics or attributes.

Again this time, QuantFM will be a Shiny app developed with the programming language R. As development is currently under progress, I can provide a provisional features roadmap for the app:

First release:
* Upload multiple squad and shortlist datasets
* Analyze squad's attributes and statistics progress if multiple squad datasets are provided
* Compute position/role/duty statistics-based and attributes-based scores and compare between any datasets - squad vs shortlist, shortlist vs shortlist, squad vs squad
* Customizable models for scores computation by the user
* Custom visualizations

Second release:
* Statistical based analysis and scouting recommendations with models trained from a set of players

Getting now to the main objective of this post - see if anyone is interested in helping out. I began development of the app less than a month ago, but between work, studies and other projects it would be difficult to ship the app in less than a month or two. In order to accelerate development, help can be provided in multiple forms, from non-coders and coders as well (the list may be updated):

• [Coding] App theming: the app's theme will be inspired by FM24 - dark mode, purple secondary background, orange accents, etc. Would need someone who knows CSS to build the style sheet.
   
• [Non-coding] Default attributes-based score models: will need to have all primary and secondary key attributes for each position/role/duty compiled as provided by SI in the game.
   
• [Non-coding] Default stats-based score models: will need to have all statistics categorized (i.e all crossing attributes, all passing attributes, etc).
   
• [Non-coding, experienced players] Custom pre-set stats-based score models: this is for experienced players, if you'd like to have your models in the app I would be happy to include them. Thus users could choose to perform their analysis with @Rashidi's presets, @Cleon's presets, etc.

Your contributions can be reflected in two ways:
* All contributors will be mentioned in the "About" section of the app with links to your SI forum profile, your GitHub or even your website
* For anyone who provides data or scripts that will be included in the app's source code, you could commit it yourself in the app's GitHub repository and be automatically credited as a contributor.

Currently, the app is in a private repository as it has not hit production yet, but for anyone who'd need to contribute in the repository access will be granted. If you can't help but are interested, feel free to leave feature suggestions in the comments! When the app will hit production, you'll be able to suggest features or reports issues/bugs directly in the app's repository.

To end this, here is a quick preview of the current version of the app - all the logic to get your datasets in the app are done and the analysis logic is under way:

 

Feel free to PM if you're interested!

Thanks

Edited November 6, 2023 by gam945

[Rashidi]

This is a very good idea. Will definitely be in touch.

[Cultured Left Foot]

Very keen to see where this goes. I'm deep into the data side of things. 

I don't have much free time, but if you need some support - I can try my best!
 

Edited November 6, 2023 by Cultured Left Foot

[gam945]

If someone would like to help with compiling the FM key and preferable attributes for each role/duty combination I'd appreciate it and would advance quicker on the app coding. The template is attached, check the Attributes sheet. Each row corresponds to the attribute and each column represents a role/duty (I have all the 90 combinations available). So basically you'd fill it column by column, the following way:

Key attribute = 1
Preferable attribute = 0.5
Other attributes: leave it blank to not waste time with inputting 0's - this can be done with the code
Do not assign anything to height, weight and left/right foot. Height may be used for center backs, goalkeepers, target mans, etc. Weight not sure but included it. Left/Right foot will be used to calculate "foot scores" and will be a bonus to the total score. For example, a player with a reasonable bad foot will have a higher score than one with a weak bad foot - all other variables being the same.

Also only fill out the role/duty combinations - generic positions (GK, FB, WB, etc) will be decided later.

 

Thanks.

models_definition_template.xlsx

[gam945]

Some progress while still on development phase - here's a preview of the Squad > Tactics page:

Still looking for people with knowledge in HTML/CSS to help me style the app..

Edited November 30, 2023 by gam945