1) There can easily be too many bonuses/penalties to manually keep track of, and there is no interface to count it up for you (at least in my version).
Making it harder to keep track of is really the point. I suppose that can be really annoying to someone that wants to maintain a master level capacity of oversight on every decision. I can understand that. Then again, I feel that if I can reach that point, the game quickly stales with the reliability of expectations.
2) Because of how the research is structured, you often end up with some amount of beelining
Causing rerouting through the tech tree is again, really a big part of the point. It adds character to the different paths that doesn't exist (at least in the same way) without the traits you've selected, forcing changes to strategy to be considered.
Because the bonuses you are interested in have the most effect in the later part of a beeline, you cannot just shift tech to test how much you bonus you would get at the point where you should be taking the strategic decision of which path to go down
You can still see, when you select to beeline on the tech tree, the total amount of turns to get to the end... I'm not sure I'm seeing how this is a problem.
Since the techs tend to be quite varied over a given beeline in type, you usually only get bonuses to a small subset of the techs in the beeline, making the gain much more marginal.
Which would mean then that if you are the type of player that just goes for what he feels is ultimately the best target to go after, then overanalyzing it beyond that point is not really necessary. Just means that you'll have a few unexpected hills and valleys along the way depending on what traits you took. If what you are going after as a priority tends to match with your playstyle, for example if you prioritize getting as much production as fast as you can and you also choose traits that enhance production, you're likely to also be, by nature of the similarity or relationships, getting to achieve those tech goals faster. If you're selecting your traits according to the overarching strategic values you bring to the game, you should be benefiting most towards making that strategy work even better for you. Part of the purpose of the complexity of the traits is to make this less about analysis and more about choosing your strengths and weaknesses and giving you the opportunity to play to your strengths and hopefully you have ideas on how to minimize or mitigate your weaknesses. The analysis tends to come naturally through experience so I'm inviting more intense and regular play experience to get you to the point of the mastery that deep familiarity brings. Once you reach that mastery, again, any game tends to stale a bit and you may go looking for other games to master.
AKA, the climb is more important than reaching the peak. If it's a short trek, it doesn't mean as much. The things you are struggling with are the very things that give the system some value. Not everything should be immediately obvious in design. Very little should be really. A great game is a box of secrets that can never be completely mined out of all its mysteries. I planted a lot of mysteries in this garden and many of them are even obfuscated from me because it will take years of playing to master a complete understanding of all the little ins and outs of my own design.
The obfuscation runs counter to the idea of varying tech paths based on traits, since that requires that the playing can see that another choise would be better in this situation. What is worse, is that the impelmentation and specifically 2b means that the obfuscation can be larger than the effect, which makes this problem even worse.
There again, when you start to gain familiarity through multiple playthroughs, you'll start feeling the landscape better. I wouldn't expect myself to even see how the changes should cause course corrections for ideal optimization of play. At least not at first. It takes a while to master a trait selection and how to make it work the best for you. Once you have that down, there are many other mountains to climb, represented by other traits and figuring out how to best play with them. It's specifically not meant to be easy to master. But once you start being able to recognize the techs that are adjusted and how that may mean a different path may be better to take for various reasons inspired by those cost differences, then you really begin to feel like you're learning something. THAT is what I want players to eventually feel, after earning it thoroughly. Such a feeling of mastery can only come from a rich system that doesn't give up its strategic depth easily.
I know I'm a little strange in that I value difficulty but to me, the very things you are saying are a problem are the very things that I feel give the system deep value.
My suggestion is to change it from changing the total research output to changing the cost of a specific research. This would stabalise the research output, thereby reducing the obfuscation, and it would be much easier to see from the research screen how a given tech path might now be cheaper/more expensive because of your traits.
This is more limiting to the outer extremes, particularly in bonuses. If you ended up building to -100% the cost, through this and possibly other sources of modification, then you'd make the tech free. Modifying the research output is much more malleable.
I get what you're saying about the averaging out of the bonuses over a long beeline since there's lots of different types of techs along that way, but it does have a greater impact for players that don't beeline as far and select according more to where they are then the far reaching goals they want to achieve. If you're beelining along far paths, then this is just a nice bonus now and then and won't really make much of a difference in your decisionmaking. And that's fine too.
C2C (and the entire Civ franchise) thrives on interesting decisions. You have the choice here of making shorter path choices to see if you can optimize your achievements by making the best use of your tech mods from traits VS just raw beelining and letting those mods be incidentals along the way.