If an ostensibly free market is not producing competition which causes that to happen, it's a problem for society; then, it is society's prerogative to use whatever means appropriate to fix it. That could include, depending on the ideology of the society, loans and grants for new participants in the market, legislation that regulates pricing, specific taxation policies, or even rejecting the free market entirely and moving to central control.
By analogy, a mysterious violent crime case might need to be solved for justice to be attained, but it doesn't need to be solved in order for a society to improve and make itself safer generally.
Some free markets result in natural monopolies so they don't always cause sales values to trend to production costs. I believe the scenario you are thinking about is "perfect competition". A free market results in perfect competition when a number of other conditions are met, including, but not limited to low barriers to entry, and lots of buyers and sellers.
No, the point of a free market is to let the price system do its thing, and prices have nothing to do with production costs as long as they are higher. If prices appear to move in unison and that actually happens to be because of collusion, in a free market a new competitor could undercut the cartel. That's assuming that the cartel members themselves don't break the collusion, which they have every incentive to.
The long-term impact of collusion on prices is limited compared with e.g. money printing and other sources of inflation.
This occurs regardless of how much competition there is.
If you call the neighboring stores and say "hey, let's all sell apples for 2x our current price so consumers don't have a choice but to buy it for that price", then you are colluding.
Your point does stand though. From the outside it's very hard to distinguish between collusion and not by mere correlation. Prices may converge for various legitimate reasons, and without insider knowledge, reliably identifying collusion is often impossible.
The best-case scenario for the project might involve comprehensively modeling a single product category to demonstrate that only collusion could explain the pricing patterns. Success would then hinge on litigation uncovering direct evidence like incriminating communications.
Did you read that on the website, or are you making that assumption?
This is just a dataset, and on its own isn't going to reveal something like collusion. But if you are researching collusion, you can use this data to help you understand behaviors of businesses.
Also, collusion doesn't necessarily have to involve prices.
If I set the price to the average, is that collusion or just trying to maximize my profit on apples?
If I set it higher than average, is that collusion, or me telling the market that I think I have premium apples?
How do you get from correlation to collusion? This project seems to build on an assumption that it is collusive to set prices based on other market prices, or else how could it be a hammer to bash collusion? Is it required to be ignorant of market prices when you set a price so that regulators can assume that correlation is evidence of an agreement?